Enter your keyword

Final Year Project

Abstracts OPEN HOUSE 2019

Introduction

The event, titled “OPEN HOUSE 2019” is aimed at providing the corporate sector with an insight into the ability and technical proficiency of the students of National University of Computer & Emerging Sciences – FAST (Karachi Campus). This event also provides a common platform for the academia and the industry to come together. Thus instigating, what may well prove to be a mutually beneficial partnership.

Director’s Message

NUCES-FAST has been striving to produce quality graduates who are able to meet real world challenges through their knowledge, wisdom, hands-on experience, problem solving abilities, drive, and professional acumen. With the active support of highly skilled and dedicated faculty, properly equipped labs, and a purpose-built infrastructure, the University produces skilled professionals in the fields of Computer Science, Electrical Engineering, and Management Sciences who are able to handle real world problems.

An important stage of evaluation and testing of innovative ideas is the Final Year Project (FYP). Students have performed extensively during the whole academic year to demonstrate results and prototypes of their efforts. These project reflects the intelligence, the hard work, the abilities, and the management skills of students. Faculty advisors also have immense contribution in bringing the outcome of these projects.

I strongly believe that these projects have the potential to solve many of our national problems. They also can be extended further for research and product development and can be adopted at national and international level.

Prof. Dr. Jawwad Ahmed Shamsi

Director

National University of Computer and Emerging Sciences (NUCES-FAST) Karachi Campus 

Open House 2019 Team’s Message

Each year, the FYP Open House fulfills FAST (NUCES)’s legacy of commitment and novelty. As the world shifts towards new technologies and tools by the minute, it is a matter of great pride for us that our graduates have kept a steady pace with these advancements and have continued to dominate the I.T. arena of Pakistan year after year.

We are very thankful to our faculty, staff and student of our university whose suggestions, cooperation, participation and continuous help has made the event a reality. We hope you find Open House 2019 intellectual stimulating and exciting.

CS FYP Committee Members

Coordinators:

Mr. Farrukh Hasan Syed

Mr. Shahbaz Siddiqui

Mr. Shoaib Raza

Members:

Mr. Muhammad Nadeem

Mr. Abdul Aziz

Ms. Farah Sadia

Mr. Basit Ali

Mr. Awais Ahmed

Mr.  Syed Zain Ul Hassan

Ms. Rahemeen

Ms. Mubashra Fayyaz

Ms. Maham Mobin

Ms. Sumaiyah Zahid

Ms. Safia

Mr. Irfan Ayub

Ms. Tooba Ali Zahid

Mr. Muhammad Fahim

Mr. Hamza Ahmed

Mr. Abdullah Somroo

Ms. Aqsa Zahid

 

EE FYP Committee Members

 

Engr.Muhammad Haris Mohsin

Engr. Qurat ul Ain Sohail

Engr. Muhammad Ahsan Khan Sunny

Engr. Zakir Hussain

Engr. Sadaf Ayesha

Dr. Muhammad Burhan Khan

Engr. Kashan Hussain (Coordinator) 

FSM FYP Committee Members

 

Mr. Syed Babar Ali

Mr. Michael Hieman Simon

Mr. Mohammad Zeeshan Ahmed

Mr. Mohammad Ahsan

Ms. Asiya Zaheer

Ms. Saima Abbas

 

 

 

 

Review Committee

Ms. Nazia Imam

Ms. Sameera Sultan

Ms. Faiza Mumtaz

 

 

 

 

 

 

Projects

Department Of Computer Science

HOME AUTOMATION USING HOLOLENS

RAPID DETECTION OF CRIME USING DEEP LEARNING ON FOG NODES

BILINGUAL CHATBOT

DYNAMIC ROUTE COMPUTATION FOR HIGH SPEED NETWORKS USING SDN

ENSEMBLE OF DEEP LEARNING AND LOCAL FEATURES FOR ROBBERY DETECTION

UNETHICAL ACTIVITY DETECTION SYSTEM

THE EYE STICK-WALKING STICK THAT SEES

P2P Real Estate System

VIOLENCE DETECTION IN CARTOONS

GUMSHUDA – FINDING MISSING PERSON USING FACE RECOGNITION

STOCK MARKET PREDICTION USING MACHINE LEARNING TECHNIQUES

REGION BASED SENTIMENTAL POLARITY IDENTIFICATION FROM ONLINE DATA OF POLITICAL AND FINANCIAL NEWS

BLOCKCHAIN TECHNOLOGY IN EDUCATION TO PREVENT FAKE DEGREES

OPTIMIZING TRANSPORTATION AND LOGISTICS USING BLOCKCHAIN (HOOK BLOCK)

SECURE E-VOTING USING BLOCKCHAIN

IMAGE PROCESSING SERVICES FOR INTELLIGENT INFORMATION SYSTEMS

FAST-COIN:BLOCKCHAIN BASED REGULARISED E-CURRENCY USING HYPERLEDGER FABRIC

FAULT DETECTION AND MONITORING OF RENEWABLE ENERGY RESOURCES

TOWARDS THE CONSTRAINT LEARNING AND OPTIMIZATION APPROACH TO DOCUMENT CLUSTERING

BUILDING DAMAGE CLASSIFIER

ADAPTIVE CLUSTERING OF SOCIAL MEDIA MESSAGES

TRAFFIC LAW VIOLATION ANALYSIS

DIGITAL TAJIR

SOCIAL SEARCHER

CROWDSOURCED REALTIME COMMUTE SERVICE

AUTISM THROUGH COMPUTING METHODS

COGIMICS – RESEARCH ON COGNITIVE ENGINE USING VISUAL, TEXTUAL AND PHYSICAL EXPRESSIONS

CRICKETER’S TALENT HUNT: A SMART RECOMMENDATION SYSTEM

SMART RECRUITANT

INTEGRATED SERVICES PROVIDING SYSTEM

SCHOOL RECOMMENDER

FYP WORK FLOW ENGINE

ENGAGEMENT PORTAL

VR SKILL SET FOR INTELLECTUAL DISABLE PEOPLE

STUDENT HEALTH EDUCATION (SHE)

SECURE COMMUNICATION FOR LOW POWERED IOT DEVICES IN NARROWBAND ENVIRONMENT

KEYNOTE SUPERIMPOSITION AND VISUAL FEEDBACK BASED PIANO LEARNING SYSTEM USING AUGMENTED REALITY

AR-GO

ADEQUATE BODY SUITING EXPERIENCE

INVESTIGATION OF MULTI-TARGET LEARNING

FOODSQUARE

DIABETIC RETINOPATHY (DR) CLASSIFICATION THROUGH SMARTPHONE FUNDUS IMAGING

DATA ANALYTICS AND CLASSIFICATION FOR ‘HYSAB KYTAB’

PERSONALIZED TREATMENT OF CANCER USING MACHINE LEARNING

APP OF KINDNESS

SMART BIKE

UNIVERSITY SMART CARD SYSTEM

JOB ASSISTANCE APPLICATION

”KAIZEN”-CARPOOL APPLICATION

3D DRAWING IN AUGMENTED REALITY

SMART SHOP

NEUROCOIN.IO

AAB GAARI

PERSONALITY IDENTIFICATION THROUGH TRANSFER LEARNING USING ARTIFICIAL NEURAL NETWORK (ANN)

QURAN COMPANION (INNOVATIVE WAY TO MEMORIZE AND REVISE QURAN)

ONLINE SKILLS, SERVICES AND BUSINESSES MARKETPLACE

PRICECOM

SMART BIDDING APPLICATION (BidX)

DRIVE LEARNING APP

MEALS ON WHEELS

PAKISTAN SIGN LANGUAGE (PSL) CHAT PROJECT

AI DOC

ECG INTERPRETATION AND RHYTHM ANALYSIS

Department Of Electrical Engineering

UNDER GROUND CABLE FAULT LOCATOR

SMART IRRIGATION SYSTEM

AUTOMATED AIR FIELD CONTROLLING AND MONITORING SYSTEM

IOT BASED SMART ENERGY GRID

IOT BASED SMART BANK SECURITY AND SURVEILLANCE SYSTEM

COMPUTER VISION BASED COMMUNICATION ASSISTANCE FOR AUDIBLY IMPAIRED PEOPLE

IOT BASED SMART MIRROR

ELECTRONIC VOTING SYSTEM

FORMATION CONTROL OF MULTIPLE UAVS

SMART ATMOSPHERIC WATER GENERATION UNIT

SMART JACKET FOR SAFETY OF COAL MINERS

AUTONOMOUS UNDERWATER VEHICLE ROBOT

Department Of Management Sciences

CONSUMER BUYING BEHAVIOR ON E COMMERCE THE CASE OF VEGETABLES

BOARD OF DIRECTORS AND FIRM’S PERFORMANCE

EFFECTS OF CAPITAL STRUCTURE ON THE PERFORMANCE OF MANUFACTURING FIRMS IN PAKISTAN

IMPACT OF RETAIL LAYOUT, DESIGN AND LOCATION ON CONSUMER BUYING BEHAVIOUR

EXPLORING STUDENT CHOICE CRITERIA FOR SELECTING BUSINESS SCHOOLS IN KARACHI

CONSUMER BUYING BEHAVIOR ON E COMMERCE THE CASE OF VEGETABLES

BOARD OF DIRECTORS AND FIRM’S PERFORMANCE

EFFECTS OF CAPITAL STRUCTURE ON THE PERFORMANCE OF MANUFACTURING FIRMS IN PAKISTAN

IMPACT OF RETAIL LAYOUT, DESIGN AND LOCATION ON CONSUMER BUYING BEHAVIOUR

EXPLORING STUDENT CHOICE CRITERIA FOR SELECTING BUSINESS SCHOOLS IN KARACHI

Department Of Computer Science

 

HOME AUTOMATION USING HOLOLENS

 

MUHAMMAD HANEEF 15k-2321 (K152321@nu.edu.pk)
ANIL KUMAR 15k-2169 (K152169@nu.edu.pk)
UZAIR RIAZ FAROOQUI 15k-2835 (K152835@nu.edu.pk)

 

SUPERVISOR

Dr. JAWWAD SHAMSI (jawwad.shamsi@nu.edu.pk)

Mr. DANISH HAMEED (supervisor@nu.edu.pk)

 

The project is based on the basic concept of Home Automation implemented using emerging technology HoloLens. Normal home appliances, such as televisions, refrigerators, fans, lights, and many others have a physical mechanical switch to turn them on/off. However, our motivation is to replace those switches with augmented 3D switches. HoloLens has unique gesture-based I/O features that will help in interacting with the physical home appliances.

HoloLens was connected with the Raspberry Pi via Internet, using Azure IoT Hub. Azure IoT Hub is a medium between both devices to exchange messages. Two Universal Windows Platform (UWP) applications were developed; one was developed for HoloLens and the other for Raspberry Pi. The Raspberry Pi is the brain which will make simple home appliances smart. We connected the bulb with the Raspberry Pi via a 4-Channel Relay and then sent a turn on/off signal to the Raspberry Pi by tapping on the virtual/augmented button visible beside an electronic device. Whenever we tapped on the button, the UWP application running on the HoloLens forwarded and stored the corresponding command at the Microsoft Azure IoT Hub and then the command sent to the connected devices at IoT Hub. Raspberry Pi supplied voltage on the basis of command turn on/off (min is 0 and max is 5). If the command received is “ON”, the bulb was successfully turned on and if the command received is “OFF”, bulb was successfully turned off.

 

RAPID DETECTION OF CRIME USING DEEP LEARNING ON FOG NODES

 

MUHAMMAD YASIR 15k-2214 (K152214@nu.edu.pk)
HAFIZA HAIQUA FATIMA 15k-2301 (K152301@nu.edu.pk)
ABDULLAH ALI MEMON 15k-2859 (K152859@nu.edu.pk)

 

SUPERVISOR

Dr. JAWWAD AHMED SHAMSI (jawwad.shamsi@nu.edu.pk)

Ms. TANIA IRAM (tania.iram@nu.edu.pk)

 

Due to the rapid urbanization, cities are converting into smart cities. Safety is an essential requirement in a smart city and the performance of the currently available surveillance systems is not real time. To solve the above mentioned problem, we propose a rapid crime detection surveillance system using fog nodes that can detect crime and take appropriate actions in real time. We propose an architecture based on fog nodes that will distribute the data constantly coming from video streams and provide storage, reduced latency, and improved processing. Although fog nodes speed up the processing, we still need a fast object detection algorithm and face recognition system that can detect criminal artifacts and criminals respectively in real time. Therefore, the robust and the most accurate object detector available to date is selected. Experiments of retraining YOLO were performed on different dataset sizes and achieved an mAP of 0.85 on Openimage test and validation set combined with Coco person dataset. We also transferred data between the fog nodes using infini band for our proposed architecture and it provides a significant reduction in transfer time between fog nodes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BILINGUAL CHATBOT

 

NABEEL ALAM 15k-2269 (k152269@nu.edu.pk)
MUHAMMAD UZAIR NADEEM 15k-2837 (k152837@nu.edu.pk)
AREEQ AZAM 15k-2322 (k152322@nu.edu.pk)

 

SUPERVISOR

Dr. JAWWAD AHMED SHAMSI (jawwad.shamsi@nu.edu.pk)

Ms. SUMAIYAH ZAHID (sumaiyah@nu.edu.pk)

 

Despite English being dominant in official communication, Pakistan is a country where Urdu is the preferred medium of communication in many scenarios. This project aims to become a bridge of communication between the general public and the health institutions of this country by providing both of them a bilingual health-care chatbot which will work just like a personal hospital/clinic representative, understanding the user’s informational needs and answering them just like a real representative would while communicating in both the English and Urdu languages. The chatbot system to be constructed in this project will be based on the Object-Oriented structure. It will have multiple components working together to constitute a whole system to achieve the required result. We are using both approaches, the probabilistic model as well as deep learning to make this chatbot more efficient. The project takes inspiration from AIML projects, such as the Chinese chatbot developed in 2014, and for the training module, it seeks ideas from researches on the Evibot, a chatbot for parents of autistic children, and a gated end-to-end memory networks based project built by Jincy Susan Thomas et. al.

 

 

DYNAMIC ROUTE COMPUTATION FOR HIGH SPEED NETWORKS USING SDN

 

MUNESH KUMAR 15k-2121 (k152121@nu.edu.pk)
SAMEER 15k-2140 (k152140@nu.edu.pk)
KASHAF MANZOOR MEMON 15k-2200 (k152200@nu.edu.pk)

 

SUPERVISOR

Dr. JAWWAD AHMED SHAMSI (jawwad.shamsi@nu.edu.pk)

Ms. NAUSHEEN SHOAIB (nausheen.shoaib@nu.edu.pk)

 

In the modern era, the world is connected by different heterogeneous links, such as Ethernet, WAN, InfiniBand, and others transferring billions of packets. The path of a data packet from the source to destination is usually based on the shortest path irrespective of connection speeds. In addition, in real world, there exists some critical data packets of applications which need to be transmitted in minimum possible time.

 

In our project, we have deployed a Software Defined Network (SDN) with heterogeneous connections having different speeds such as Ethernet and InfiniBand. We have achieved the speed of around 51 Gbps using RDMA over InfiniBand which is capable to meet requirements for large data centers. We will implement the routing framework algorithm that tries to find the route with a high speed connection considering the priority of a packet. We will focus on minimizing the response time as our metric for packet transmission.

 

Big Data is one of the biggest constraints in today’s network environment which requires high speed connection links, such as InfiniBand. This architecture will help high priority packets to be transferred in minimum possible time and hence, can easily be deployed in large data centers and organizations to manage the huge amount of data more efficiently.

 

 

 

BILINGUAL AUTOMATIC SPEECH RECOGNITION SYSTEM

 

DANISH MAHMOOD MALHI 15k-2134 (K152134@nu.edu.pk)
AMEER HAMZA RIAZ 15k-2809 (K152809@nu.edu.pk)
AKRAM HANIF 15k-2270 (K152270@nu.edu.pk)

 

SUPERVISOR

Dr. JAWWAD AHMED SHAMSI  (jawwad.shamsi@nu.edu.pk)

Mr. DANISH HAMEED  (supervisor@nu.edu.pk)

 

We aim to create a Bilingual Automatic Speech Recognition system for Urdu speakers who frequently tend to use English words and sentences in between their Urdu speech. The system is needed because the English language has extensive ramifications in the Urdu language for the common Urdu speaking person. Our ASR will be able to detect and recognize both pure Urdu and English sentences as well as Urdu sentences with English words mixed in and produce their transcripts respectively.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ENSEMBLE OF DEEP LEARNING AND LOCAL FEATURES FOR ROBBERY DETECTION

 

JUNAID WAHID 15k-2238 (K152238@nu.edu.pk)
SARVECH SAMEJO 15k-3615 (K153615@nu.edu.pk)
YOUSUF TANVIR 15k-2116 (K152116@nu.edu.pk)
RIDA ALI WAQAS 15k-2221 (K152221@nu.edu.pk)

 

SUPERVISOR

Dr. ATIF TAHIR (atif.tahir@nu.edu.pk)

 

In our daily lives, we all witness or hear about robberies every other day, but rarely do we hear that the law enforcement officials were notified during the time of the act and/or they stopped it. It is only after the robber has robbed and left the place that the personnel are free to move and alert the local law enforcements. CCTV camera, however, watches and usually records the entire act. Usually, there is a person who is monitoring the CCTV feed which can then detect the robbery. Nevertheless, this method is not only expensive but can fail due to human error. For example, the guard can fall asleep, go for a break, be bribed, or even get hurt by the robbers.

Our project aims to automate this process by detecting robbery using just the CCTV video feed as input. Frames are extracted from the input video and are processed in our detection module which is build using a combination of three techniques of computer vision and image processing, which then provide an output in the form of a binary classification of whether there is robbery or not.

 

 

MISSING PERSONS RECOGNITION USINGAGE INVARIANT FACE RECOGNITION

 

SAAD BIN SHAMEEM 15k-2130 (k152130@nu.edu.pk)
ADEEL TAHIR RANA 15k-2178 (k152178@nu.edu.pk)
MUHAMMAD UMER WAJIH 15k-2883 (k152883@nu.edu.pk)

 

SUPERVISOR

Dr. ATIF TAHIR (atif.tahir@nu.edu.pk)

 

One of the most challenging tasks in face recognition is achieving age invariance. The project focuses on developing a system which can detect the face of a child or an adult person which is invariable of changes that occur with age. Our project will use facial recognition using cross-age referencing to detect the identity of that person. The objective of this project is to identify missing people that may have their facial features changed overtime due to age. This project addresses a problem of identifying a person who has been missing for a long time and their facial features may have evolved with age. This is most common in children as when they grow, their face changes its features..

 

 

 

 

UNETHICAL ACTIVITY DETECTION SYSTEM

 

S.M KHIZER AHMED 15k-2136 (k152136@nu.edu.pk)
M. FAIZAN MAZHAR 15k-2366 (k152366@nu.edu.pk)
KARIM BUKHSH ARBAB 15k-2199 (k152199@nu.edu.pk)

 

SUPERVISOR

Dr. ATIF TAHIR  (atif.tahir@nu.edu.pk)

 

It is a well-known fact that in our region specifically, Pakistan, the rate of people spitting, littering, and engaging in other indecent activities is very high. We have targeted such offenders in our project and the motivation is to basically reduce these problems. First of all, the system will take in input video stream from the surveillance cameras and pass it on to the software which in turn would break the input video stream into RGB for further processing. The next few steps involve generating motion influence map for frames to represent the interactions that are captured in a frame. The main characteristic feature of the proposed motion influence map is that it effectively depicts the motion characteristics of the movement speed, movement direction, and the size of the objects, and their interactions within a frame sequence. It further extracts frames of high motion influence values and compares them with the testing frames to automatically detect unusual activities. After any anomaly is found, the system would take an image of that action and would send it to the concerned authorities via desktop application interface.

 

 

 

HARD DISK FAILURE PREDICTION

IN LARGE DATACENTERS

 

MUHAMMAD ANAS   15k-2191 (k152191@nu.edu.pk)
MUHAMMAD SAFI   15k-2143 (k152143@nu.edu.pk)
WALEED HASAN   15k-2193 (k152193@nu.edu.pk)
MUHAMMAD NOUMAN   15k-2240 (k152240@nu.edu.pk)

 

SUPERVISOR

Dr. ZULFIQAR ALI MEMON (zulfiqar.memon@nu.edu.pk)

 

The failure of hard disks in large datacenters and the unscheduled maintenance time to fix them have often made a high magnitude of servers unavailable and can increase the service latency. We will propose a method to predict failure in a server’s hard disks thereby making datacenters more reliable and fault tolerant. Our prediction algorithm will focus on improving the stability and predictability in large datacenters using machine learning with a presentable amount of accuracy to increase the efficiency and decrease the downtime of data centers.

 

 

 

THE EYE STICK-WALKING STICK THAT SEES

 

TALHA AHMED SHAIKH 15k-2202 (k152202@nu.edu.pk)
ABDULLAH SIRAJ 15k-2211 (k152211@nu.edu.pk)
MAHNOOR IFTEKHAR 15k-2222 (k152222@nu.edu.pk)
RAWAHA BIN KHALID 15k-2879 (k152879@nu.edu.pk)

 

SUPERVISOR

Dr. ZULFIQAR ALI MEMON (zulfiqar.memon@nu.edu.pk)

 

 

According to the US Census Bureau Report 2012, 56.7 million people in this world which is almost 19% of the world’s population, are suffering from some sort of physical impairment that affects their daily life. This 19% population of the world requires some sort of external support to perform the daily chores of their life. As the technology has progressed, various companies, such as Google, Indiegogo, and Microsoft have produced few assistive gadgets, some of them are Google Glass and WeWalk to create an impact, but these tools and instruments are quite costly for the majority of the affected class. Over 3 billion people in this world are living a life of poverty and cannot afford the apparatus that eases their lives. The purpose of this project is to produce a prototype of a cost effective and affordable solution for the people who face troubles due to some sort of physical disability. The Eye Stick with an infra-red, pressure, and moisture sensor along with a camera together with bluetooth connectivity with a back end neural network engine is being developed considering the latest technological trends and social requirements. Currently the hardware device with embedded sensors is functional. An object classification module that classifies few basic objects is also to be embedded in the final product to assist the person in navigating through the paths. The most advanced feature planned for this is real time image processing with voice command recognition. 65% of the product is ready, but rigorous testing of features is required considering the real world environment.

 

 

DEAF’S SOLUTION

 

KHYZER MANZOOR 15k-2874 (k152874@nu.edu.pk)
ASFAND YAR  15k-2104 (k152104@nu.edu.pk)
SHERYAR ALI 15k-2141 (k152141@nu.edu.pk)
ROHIT KUMAR 15k-2273 (k152273@nu.edu.pk)

 

 

Supervisor

Dr. ZULFIQAR ALI MEMON (zulfiqar.memon@nu.edu.pk)

 

More than 5 percent of the entire world population, that is 360 million people, is deaf. This disability poses significant problems for the deaf when communicating via phone. To facilitate such people, this project proposes an application call Deaf’s solution. Deaf’s solution will integrate hearing impaired community around the world into the mainstream of society by providing a platform which will bridge the gap between deaf and non-deaf person. The concept is simple and concrete where the speech of live call is converted into text and if someone is Deaf and dumd, for them this application will convert text into speech.

The idea of this project is to develop a mobile Application. This Application consist of two main phases. The first phase is Speech to text and second phase is text to speech for users who are Deaf and Deaf and Dumb both.

 

 

Learning Imbalanced Datasets using Differentiated Regularization of the Loss Function

 

BAHRAM BALOCH  15k-2817 (k152817@nu.edu.pk)
ABEERAH REHMAN  15k-2840 (k152840@nu.edu.pk)
SANJAY HARESH  15k-2179 (k152179@nu.edu.pk)
SATEESH KUMAR 15k-2821 (k152821@nu.edu.pk)

 

SUPERVISOR

Dr. TAHIR Q. SYED (tahir.syed@nu.edu.pk)

 

The class imbalance problem is common in most real-world machine learning problems. Generally, it is mitigated using sampling methods, classifier ensembles or cost-sensitive learning. In this work, we propose a novel loss function, the Focused Anchors Loss, by combining cost-sensitive learning with discriminative feature learning. We evaluate our loss function on 6 highly-imbalanced datasets and the experiments show that the proposed loss function and its variants consistently outperform cross-entropy loss and focal loss, and produces comparable or better results to sampling methods without having to incur the extra computational cost associated with those methods.

 

 

 

P2P Real Estate System

 

SYED ZAIN HUSSAIN  15k-2213 (k152213@nu.edu.pk)
MUHAMMAD RIZWAN JAVAID 15k-2162 (k152162@nu.edu.pk
ONEED IQBAL 15k-2862 (k152862@nu.edu.pk)
MALIHA KHIZER 15k-2866 (k152866@nu.edu.pk)

 

SUPERVISOR

Dr. ZEESHAN AHMED (zeeshan.ahmed@nu.edu.pk)

 

The aim of the project 2 is to maintain a real estate record on blockchain through which buying and selling can be possible via an online web application and mobile application. Furthermore, the application will provide a portal for buyers and sellers to buy and sell through smart contracts. This will ensure the transparency as the records will be maintained on the blockchain.

 

 

VIOLENCE DETECTION IN CARTOONS

 

RIMSHA MAJEED 15k-2234 (k152234@nu.edu.pk)
FARAZ AHMED 15k-3616 (k153616@nu.edu.pk)
AHMED BILAL 15k-2223 (k152223@nu.edu.pk)

 

SUPERVISOR

Dr. ZEESHAN AHMED (zeeshanahmed@nu.edu.pk)

Dr. ATIF TAHIR  (atif.tahir@nu.edu.pk)

 

Cartoons are an informative way for creating awareness; children take keen interest in watching cartoons and spend leisure time in front of television. Unfortunately, there is an increasing trend of violence and other objectionable scenes in cartoon videos that have bad impact on the personality of the youth. Extensive use of such violent scenes in media (movie and TV programs) is one of the factors responsible for increasing violence in the society. In this paper, we try to solve this problem by suggesting a novel technique of classifying a video as violent/non-violent by training Convolutional Neural Network (CNN) on the cartoon dataset.

 

We have made our own dataset consisting of violent and non-violent videos. The data are in the video and the model is fed images that are the frames extracted from the video data with a total image count of about 38,449 images with 11,169 non-violent and 27,280 violent images in separate folders marked as labels before preprocessing kicks in. We have selected the intensities of the images as our features for the neural network. Initially, we used a vectored version of all images which was of the size (320 x 240) which was humongous along with our dataset and the network when training dropped into the problems of memory issues on a machine with fairly good specifications. Thus we took the decision of skewing the data at hand by converting each image used to a size of (64 x 48) and the images were then saved for further processing.

 

Our model produced results with a classification rate of 69.07% where the number of epochs was equal to 100 with a dropout rate of 0.5. In terms of future work, the idea is to extend these results further and achieve better accuracy by tweaking hyper parameters and experimenting with the network architecture.

 

 

 

GUMSHUDA – FINDING MISSING PERSON USING FACE RECOGNITION

 

NISHA DEVI  15k-2314 (k152314@nu.edu.pk)
ROHAT SAGAR 15k-2309 (k152309@nu.edu.pk)
AVEENASH 15k-2101 (k152101@nu.edu.pk)

 

SUPERVISOR

Dr. ZEESHAN AHMED (zeeshanahmed@nu.edu.pk)

 

 

In the field of computer vision, facial recognition is considered one of the most breakthrough advancements. Facial recognition is changing the lives of people in every aspect. One of the aspects where facial recognition can be used is recognizing missing persons. Missing persons’ problem is a big issue worldwide. An average case takes about 2 to 3 months to be solved and the risk of kidnapping and ransom increases day by day. Hence, a solution of a specified web portal with a facial recognition module is needed where posts related to missing and found persons can be shared and a platform for missing persons especially children could be made.

STOCK MARKET PREDICTION USING MACHINE LEARNING TECHNIQUES

 

RIJA FAROOQI 15k-2170 (K152170@nu.edu.pk)
                                             HANI BATOOL 15k-2819 (K152819@nu.edu.pk)
TUBA HAIDER 15k-2841 (K152841@nu.edu.pk)

 

SUPERVISOR

Dr. ZEESHAN AHMED (zeeshan.ahmed@nu.edu.pk)

 

Stock market prediction has been an issue for several years and has been one of the most significant topic of interest for researchers. The stock market is an important and active part of today’s financial markets, and both investors and speculators would like to make better profits by analyzing relevant market information. Predictions have been made on several basis and one major source of market information has been news articles. Bag-of-words based model approaches news articles by vector space model which translates each news piece into a vector of word statistical measurements, such as the number of occurrences etc. Machine learning models are then employed to capture the relationship between the word statistical patterns and the stock price movements. In our research project, we will be using different machine learning techniques to predict the stock price movement. One of the other approaches is sentiment analysis which improves prediction accuracy. We aim to use natural processing techniques to predict stock market price and using different deep learning models to achieve results.

 

 

REGION BASED SENTIMENTAL POLARITY IDENTIFICATION FROM ONLINE DATA OF POLITICAL AND FINANCIAL NEWS.

 

MUHAMMAD AHMAD KHAN  15k-2886 (k152886@nu.edu.pk )
AHMED JAFFER  15k-3631 (k153631@nu.edu.pk )
TALHA MUSTAFA  15k-3618 (k153618@nu.edu.pk )
MUHAMMAD HAMZA  15k-3619 (k153619@nu.edu.pk )

 

SUPERVISOR

Dr. ZEESHAN AHMED (zeeshanahmed@nu.edu.pk)

 

Currently, many newspapers are published online. Some of them are published dedicatedly as online editions while others are published as print edition in PDF or similar formats. In addition to newspapers, there are a wide range of opinion-oriented news articles posted online by journalists on news websites. This opens up the possibility of detecting regional sentiment for certain news by analyzing news articles which are published online on the regional news websites. These opinions are directly related to what people of that region feel regarding certain news/event and these feelings can be classified as positive, negative or neutral. This can then be plotted on a map (by writing an interface to Google maps) with different colors showing different sentiments from different parts of the world towards the same event. These sentiments (if combined from different regions of the world) can be used for many purposes especially for decision making and giving concrete information. Hence, if we combine the results of sentiments regarding certain news from different regions, we can correctly classify the emotional reaction of people around the globe pertaining to that news story.

 

 

BLOCKCHAIN TECHNOLOGY IN EDUCATION TO PREVENT FAKE DEGREES

 

M.WADEED SIDDIQUI 15k-2339 (K152339@nu.edu.pk)
MUNEEB ASIM 15k-2900 (K152900@nu.edu.pk)
ABDUL KARIM KHAN 15k-2144 (K152144@nu.edu.pk)

 

SUPERVISOR

Dr. SUFIAN HAMEED  (sufian.hameed@nu.edu.pk)

 

 

A degree is an important part of an individual’s career. It is the landmark that is achieved after

completing four years of education in a particular field. During these four years, a student has

to complete a set of courses, projects, and maintain a certain CGPA according to the standards provided by the HEC (Higher Education Commission, Pakistan). The Degree is awarded by the educational institution after authorization form the HEC. The rights are given when the institutions meet the standards and follow the guidelines provided by the HEC. The institution is also responsible for the authentication and verification of the Degree it is awarding. In order to verify the degree, the concerned institution is contacted.

The problem in this system is that an institution can obtain the license from HEC through bribery or simply forge documents. Afterwards, the institution can award Degrees or sell them to anyone and will not be caught until a deep investigation takes place. A person

having a fake degree can apply for a job and get his/her degree verified from the fraudulent institution and escape accountability until he/she is caught.

We have proposed a solution using the block chain technology through which we will have a dis-

tribute system consisting of the HEC, educational institution, companies, and the people possessing degrees. As the block chain is fault or tamper proof, any non-verified person cannot intervene in the system. Any document that is forged will easily be detected through the calculation of hash. Through our web and mobile application, a user can easily add his/her documents to the system and if the documents are original, they will be pushed in the block chain and if not, they will be not allowed to enter the system.

 

 

 

PREVENTING COUNTERFEIT DRUGS USING BLOCKCHAINS IN PHARMACEUTICAL INDUSTRY SUPPLY CHAIN

 

MEER ASIM ALI BHAYO   15k-2857 (K152857@nu.edu.pk)
MUHAMMAD MOIZ ARIF 15k-2146 (K152146@nu.edu.pk)
KHALID AHMED  15k-2803 (K152803@nu.edu.pk)
ISHTIAQUE ALI  15k-2215 (K152215@nu.edu.pk)

 

 

SUPERVISOR

Dr. SUFIAN HAMEED (sufian.hameed@nu.edu.pk)

 

The current pharmaceutical supply chain has many major issues that lead to counterfeit drugs to be distributed. Moreover, counterfeit drugs have enabled an illegal trade industry worth billions of dollars which the authorities are unable to stop, the main reason being the wide communication gap between the stakeholders of the Pharmaceutical Supply Chain. After the drugs are produced by manufacturers, they enter the pharmaceutical supply chain where they change ownership quite frequently from manufacturers to distributors to pharmacies to customers. As these drugs change ownership frequently, it is hard to say with certainty that these drugs are the same as the drugs produced by the manufactures. A possible solution for this problem is the use of private permissioned block chains which can be used to ensure if the drugs came from legitimate sources. Different stakeholders will use this blockchain to distribute the drugs further down the chain. This would enable traceability, visibility, and security to the supply chain. The biggest benefit of this solution will be that it will enable the government agencies to properly monitor the supply chain of the drugs, and be notified of any discrepancies in the changing of ownership of drugs to take legal actions, if appropriate, against the perpetrators.

 

 

 

OPTIMIZING TRANSPORTATION AND LOGISTICS USING BLOCKCHAIN (HOOK BLOCK)

 

QAMAR-UL-ARFEEN 15k-2180 (k152180@nu.edu.pk)
ARHAM ASHRAF 15k-2173 (k152170@nu.edu.pk)
ABDUL SABOOR 15k-2267 (k152267@nu.edu.pk)
HASAN ILYAS 15k-2236 (k152236@nu.edu.pk)

 

SUPERVISOR

Dr. SUFIAN HAMEED                        sufian.hameed@nu.edu.pk

 

Central Idea of the project:

 

In the age of advancement and technology the modern-day networking and transaction of goods are also moving towards the faster and better alternative the internet. The supply chain networks are also shifting from its conventional registering style to the internet. The main idea behind development of this project is to maintain a smooth and transparent supply chain system which can maintain the ledger and transaction of the fisheries in Block Chain.

 

Basic Workflow of the System

 

Our system works on the basic structure of the supply chain while our ledger of records is generating and maintaining over blockchain. The supply chain structure has three stakeholders.

  1. Supplier (Fishermen)

The supplier has the functionalities of inserting the raw fish in its database and then it can process those raw fishes into finished products and then maintain its stock. Supplier can also check incoming orders via user interfaces provided by the application.

  1. Distributor

The distributor is working as the middle man in this supply chain which is basically getting the fish from the supplier and then distributing it to the customers and other entities.

  1. Customer

The customers have the facility to order the fish from the distributor and easily get the most authentic fish on their doorstep. Because our project is maintaining and checking the authenticity of any product.

  1. Admin (Regulatory authority)

The admin has the duty of registering the valid and most authentic suppliers in the system so that the issue of any fake product in the supply chain will reduce to its minimal. It also maintains the most effective and reasonable prices of the products so there will be no major fluctuations in the prices of the fish and the taxation and maintaining proper channel in the system is also a part of duties for the Admin. The admin also checks that the supplier’s catch is from the territory from where he belongs or else their lot will not be accepted for the chain.

 

Conclusion

 

This system can be a game changer in the industry of fish and seafood by proper regulating the valid fish and reducing the quantity of black-market products in the market. The block chain technology ensures that the final products are valid and are caught by the providers by all valid means and also it did not harm the wildlife by overfishing.

 

 

 

SECURE E-VOTING USING BLOCKCHAIN

 

MUHAMMAD SHAHZAR KHALID 15k-2805 (k152805@nu.edu.pk)
ABDUL MUNIM KHAN15k-2897 (k152897@nu.edu.pk)
ABDUL BASIT 15k-2244 (k152244@nu.edu.pk)

 

SUPERVISOR

Dr. SUFIAN HAMEED (sufian.hameed@nu.edu.pk)

 

Using BlockChain concepts we can make the voting process secure and free from any malevolent changes. The e-voting BlockChain System will also be able to run on distributed systems in the future. Currently, it is centralized as ECP. The e-voting BlockChain System will have control and will store the BlockChain ledger and make the votes incorruptible. No changes can be made after a vote has been cast. It would be impossible to overrun or potentially attack the system as the Byzantine Generals problem; 51% of network majority would be needed to mutate data. By using the blockchain concept, we have developed our own permissioned blockchain. Since it is homebrewed, it is being developed to suit our needs and will accompany any design changes to the blockchain concept if required. We will try to develop the Blockchain so it can be ported to other frameworks in the future if need be. Our Blockchain will have 4 roles, an ECP which will release the initial votes, the Voter, the Presiding Officer, and our verified Vote bank. The votes will then be transferred in the same order, as cryptocurrency of a kind which will make the Blockchain immutable and none double spendable. Our RESTful API will handle the blockchain and both our interfaces (Android and Web).

 

 

IMAGE PROCESSING SERVICES FOR INTELLIGENT INFORMATION SYSTEMS

 

IBADULLAH KHAN 15K-2122 (K152122@nu.edu.pk)
MUHAMMAD ADEEL SHAH 15K-2174 (K152174@nu.edu.pk)
USAMA BIN ABUBAKAR 15K-2137 (K152137@nu.edu.pk)
MUHAMMAD 15K-2231 (K152231@nu.edu.pk)

 

SUPERVISOR

Dr. FAROOQUE HASSAN KUMBHAR(farooque.hassan@nu.edu.pk)

 

In the contemporary world, urban mobility is one of the known challenges to be tackled in the administration of a big city. This project aims to implement a software model which will provide different types of services to a client. The software model will use image processing techniques for detection of vehicles, license plate, etc. This model will then be used to provide different web services for authorized government officials. The proposed model is a combination of the latest and efficient algorithms designed specifically for densely populated cities. Different applications such as automated toll payment system and criminal identification will be implemented. In addition, several traffic measures, including, traffic flow, occupancy, and density are estimated based on the analysis of real time image processing from surveillance cameras. The availability and reliability of these traffic measures provide critical information for monitoring transportation and intelligent traffic control. Mobile applications will be provided to the end users for the monitoring of their respective activities.

 

 

 

FAST-COIN:BLOCKCHAIN BASED REGULARISED E-CURRENCY USING HYPERLEDGER FABRIC

 

MOHAMMAD YOUSUF BIN ASIM  15k-2285                                  (k152285@nu.edu.pk)
ABDUL QADIR GHAFFAR 15k-2217              (k152217@nu.edu.pk)
SYED MOHAMMAD NAEEM 15k-2864              (k152864@nu.edu.pk)

 

SUPERVISOR

Dr. FAROOQUE HASSAN KUMBHAR  (farooque.hassan@nu.edu.pk)

 

In an online payment system, people transfer digital currency from one account to another account. However, due to certain security concerns, many people do not prefer to use online payment systems and use conventional payment systems. We aim to solve such security issues by developing a project which will be able to provide a full trust worthy environment to the user and will encourage the users to use online payment systems rather than conventional systems. The project which we are developing is the “Regularized Digital Currency Using Blockchain with Hyperledger Technology”. This project is based on block chain which is used for providing security to the payment system. Block chain is the chain of different blocks which are maintained in a distributed network due to which whenever a user makes online payment, which is called as transaction, a block is created for that transaction and the transaction is stored in block after it has been validated from all of the nodes in the network. Therefore, in such a system, there is no chance of any fraud or any discrimination and so it provides full secure environment. In our project, we are going to develop a digital coin using this block chain network and a mobile application for the transactions of these coins

 

 

 

 

 

 

 

 

FAULT DETECTION AND MONITORING OF RENEWABLE ENERGY RESOURCES

 

SHUJAAT ALI SHAH BUKHARI 15K2858 (k152858@nu.edu.pk)
KUNBH KARAN 15K-2127 (k152127@nu.edu.pk)
MUHAMMAD JUMAN 15K-2203 (k152203@nu.edu.pk)

 

SUPERVISOR

Dr. FAROOQUE HASSAN KUMBHAR (farooque.hassan@nu.edu.pk)

 

Electricity consumption will comprise an increasing share of global energy demand during the next two decades. In recent years, the increasing prices of fossil fuels and concerns about the environmental consequences of greenhouse gas emissions have renewed the interest in the development of alternative energy resources.

 

The potential of renewable energy sources is enormous as they can in principle meet many times the world’s energy demand. Renewable energy sources such as biomass, wind, solar, hydropower, and geothermal can provide sustainable energy services based on the use of routinely available, indigenous resources.

 

As the demand of renewable energy source continues to grow, reducing the operational and maintenance costs and reliability have become top priorities in Wind Turbine Maintenance strategies. Maintaining a reliable operation of the system requires that the crucial components are monitored on a continuous basis and timely alarms are sent to operators. For instance, power transformers are in particular critical to this mission. Any minor failure, such as a fluid leak or progressive internal insulation degradation in transformers may threaten the safe and reliable grid operation. One study reported that a catastrophic failure in a bushing rod of a high voltage transformer costs more than three million dollars, but such a failure could be repaired easily for as little as sixteen thousand dollars if addressed in a timely fashion.

 

The advanced monitoring and indication systems that we offer provide the opportunities to anticipate, detect, and respond rapidly to sustained and/or impending failures resulting in a substantial reduction of failure rates, repair times, maintenance costs, and risk of cascading blackouts. The system comprises of the hardware, software, analytics, database, APIs, and decision-making which keeps track of abrupt changes, alarming conditions, performance, and quality assurance. The raw data is being retrieved using REST API through sensors installed at the site; normalization and meaningful data is also being taken out of the Raw Data for Visualizations, modeling, decision-making, and reasoning purposes to achieve the objectives including productivity, efficiency, condition monitoring, reliability, and security. Timely alarms and Notifications are to be sent upon any alarming situation on our Web and Application; also, overall conditional monitoring can be done remotely on Web and Mobile application. Consequently, making a two-way communication between the system and the operator helps to achieve the core requirements of the user.

 

 

DOCUMENT CLUSTERING BY CHARACTER-LEVEL FEATURE LEARNING THROUGH DEEP LEARNING

 

SYED HAMMAD JAFFERY 15k-2890 (k152890@nu.edu.pk)
ABDUL MOIZ ZAVERI 15k-2854 (k152854@nu.edu.pk)
ATHAR ABBAS 15k-2839 (k152839@nu.edu.pk)
ARSALAN 15k-2235 (k152235@nu.edu.pk)

 

SUPERVISOR

Dr. MUHAMMAD RAFI (muhammad.rafi@nu.edu.pk)

Ms. NIDA PERVAIZ (nida.pervaiz@nu.edu.pk)

 

 

Text representation learning is vital for many text mining applications specially for text clustering task. Document clustering is a challenging unsupervised problem which tries to create succinct sub-groups of documents from a heterogeneous collection of documents. The objective of the clustering task is to produce high intra-cluster similarity and low inter-cluster similarity.  Implicitly identifying patterns from textual documents for clustering is important. This Project introduces a unique character level pattern learning for document clustering. This project proposes a novel character-level autonomous feature learning and clustering approach based on neural processing. The algorithm works on two distinct steps (i) the features set for document are learnt through a character-level convolutional neural network, and (ii) the clustering is performed again using the feature based neural network using deep embedding clustering approach. The proposed approach is evaluated on standard text mining datasets using unsupervised accuracy as the evaluation measure. Encouraging results are obtained.

 

 

TOWARDS THE CONSTRAINT LEARNING AND OPTIMIZATION APPROACH TO DOCUMENT CLUSTERING

 

FIZZA ABID 15k-2853 (K152853@nu.edu.pk)
HAMZA MUSTAFA KHAN 15k-2832 (K152832@nu.edu.pk)
ANUM MIRZA 15k-2870 (K152870@nu.edu.pk)

 

SUPERVISOR

Dr. MUHAMMAD RAFI (muhammadrafi@nu.edu.pk)

 

This Final Year project (FYP) proposes autonomous constraints learning from a document collection to incorporate these constraints into an effective document clustering. The semi-supervised approach to document clustering where some prior knowledge about the collection is readily available has proven effectiveness. The project proposes algorithms based on sampling to find three different kind of constraints from the document collection: (i) instance level (ii) cluster level and (iii) corpus level. This algorithm is integrated into constraint K-Mean to produced multiple clusters satisfying the constraints.  Instance level constraints are learnt using similarity measures. Cluster level constraints are learnt by utilizing the elbow method. The Corpus level constraints are learnt from finding semantic relation between lexical features and subset of corpus. The learnt constraints produced better clustering results in both, quality of clustering and time of convergence to some extent. A boosting method is proposed to adaptively learn the constraint’s priorities, constraint satisfaction, and find an optimal clustering solution from multiple possible solutions. The project describes the design, implementation, and test of all algorithms to achieve the best clustering results on standard document clustering datasets and evaluation measures through a series of experimental studies.

 

BUILDING DAMAGE CLASSIFIER

 

MUHAMMAD ALI ZIA 15k-2123 k152123@nu.edu.pk
SAAD KAMRAN 15k-2206 k152206@nu.edu.pk
SARIM BALKHI 15k-2828 k152828@nu.edu.pk

 

SUPERVISOR

Dr. MUHAMMAD RAFI (muhammad.rafi@nu.edu.pk)

 

               

The emergence of social media sites especially Twitter has been considered a useful platform for communication and is widely used for disseminating real time information specifically during crisis such as natural disasters, but the sheer volume of the data and its sparsity makes it difficult for disaster-affected communities and professional emergency responders to process the information in a timely manner. Furthermore, the usefulness of such tweets varies as they can be very subjective at times and obtaining effective situational awareness from such images and text of the tweets becomes a challenge. The FYP extrapolate the data in order to aid relief efforts by providing actionable information about the damages caused by disasters. The FYP uses multiple methodologies that are currently dominant in natural language processing and image classification and finally presents a hybrid approach whereby an ensemble model is presented with natural language processing model and image classification model is used to solve this problem. Among the challenges faced in this regards to text was that the language used in tweets tend to vary based on region and do not follow grammatical rules. Therefore, conventional natural language processing techniques do not yield good results. The project was tuned towards Convolutional Neural Network(CNN) with embedding from text and image. The suggested model achieved 95% accuracy. The biggest problem in regard to image classification was that a lot of irrelevant images are coupled with tweets so a two pronged approach was implemented whereby irrelevant images were removed and then classification is done in terms of severity of damage.

 

 

ADAPTIVE CLUSTERING OF SOCIAL MEDIA MESSAGES

 

ABDULLAH 15k-2242 (K152242@nu.edu.pk)
ASFAND ALI 15k-2154 (K152154@nu.edu.pk)
HAMID MAHMOOD 15k-2205 (K152205@nu.edu.pk)

 

SUPERVISOR

Dr. MUHAMMAD RAFI (muhammad.rafi@nu.edu.pk)

 

Social Media websites generate a lot of textual messages related to situational news and events. These massive messages form a stream of text are often combined with tags and keywords. A target application of such highly interactive social communication is to identify clusters of events. These clusters may be used to track many demanding tasks such as engineered misinformation campaigns, political and social events, disastrous incidents, and spontaneous communications.  This is quite challenging as the data is continuously coming with new events.  This FYP proposes a novel text stream clustering approach using adaptive text features, network features, and meta data of social network. The project first collects a data stream from SOCOL platform and social media websites, and then designs and implements text stream clustering algorithm and experimentally evaluates the effectiveness of the algorithm through both internal and external measure of cluster evaluation. Topic detection model is employed to identify changing topisc and an adaptive clustering stream is deployed to re-cluster the stream.

 

Keywords: Text Stream, Clustering, experimental evaluation, text features adaptation, representational learning

INTELLIGENT BIG DATA MANAGEMENT AND STUDY CONSOLE

 

ASHAR ALI 15k-2172 (K152172@nu.edu.pk)
SAIM MALIK15k-2115 (K152115@nu.edu.pk)
SYED ROSHAAN 15k-2867

WAQAR HUSSAIN 15k-2894

(K152867@nu.edu.pk)

(K152894@nu.edu.pk)

 

 

SUPERVISOR

Dr. MUHAMMAD RAFI (muhammad.rafi@nu.edu.pk)

 

There are many driving factors for the fast growth and acceptance of cloud provision in the organizations. The cloud offers minimization of up-front IT Infrastructure cost, rapid development and deployment of applications, improved manageability, and maintenance of the resources. The disparity in cloud services and tariff compels organizations to migrate from local-infrastructures to cloud, cloud to cloud, or multi-clouds. The variety of data, such as structured, semi-structured, and unstructured formed with high velocity of generation add more challenges to the migration process. This Final Year Project proposes an Intelligent Big Data Management and Study Console (IBDMSC). The idea is to provide an integrated application that helps in migration, management, monitoring, and archiving local, cloud, and multi-cloud big-data repositories. This application will demonstrate the migration process with big-data sets from one-host to another host along with validation, logging, and archiving. This application also enables time estimates and storage analysis for the migration process. The project proposed a layer model for the application, testing, and deployment to local, cloud, and multi-cloud platform and benchmarking the migration process for some standard big-datasets of structured, semi-structured, and unstructured data in such an environment.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TRAFFIC LAW VIOLATION ANALYSIS

 

MUZAMMIL AHMED KHAN 15k-2167 (K152167@nu.edu.pk)
ASHOK KUMAR                      15k-2292 (K152192@nu.edu.pk)
PARDEEP KUMAR                  15k-3624 (K153624@nu.edu.pk)
GHULAM MUSTAFA              15k-2192 (K152192@nu.edu.pk)

 

SUPERVISOR

Dr. NOUMAN DURRANI (muhammad.nouman@nu.edu.pk)

Dr. M ATIF TAHIR           (atif.tahir@nu.edu.pk)

 

Pakistan is one of the countries that has very poor enforcement of traffic laws resulting from many factors, such as heavy traffic, political factors, corruption of traffic police, and many others. According to World Health Organization (WHO), more than twenty-five thousand people are killed on Pakistani highways annually and enforcement of traffic laws falls at number two on band of 1-10(WHO). This poor system also affects the generation of revenue by not ticketing the cars which have broken any traffic law, or have not paid tax. In this project, we have developed a solution for the above-mentioned problems with the help of computer vision and machine learning. By using image processing and pattern matching, we identify the car plate of the car via cameras while the car is passing through the major points of roads. By identifying the car plate, we check the details about the car which are whether it is a valid car plate number or not, whether it has paid tax or not, or any other dues remaining and whether this car is suspended or not. With this identification, further action can be taken against the defaulter according to the law and this will in turn assist in tax collection. It also identifies if the car is parked in no parking area and over speeding. We also track a stolen car by using these cameras. If such a car is seen in any camera,  it generates a message to the relevant authorities about the current location of the car.

 

 

REAL TIME FACE DETECTION

 

 

FOHAIL MOHAMMAD                 15K-2112

AREEBA MALIK                            14K-2043

(k152112@nu.edu.pk)

(k142043@nu.edu.pk)

MOHAMMAD SOHAIL KHATRI 15K-2119 (k152119@nu.edu.pk)
RASHID AHMED ABBASI            15K-2230 (k152230@nu.edu.pk)

SUPERVISOR

Dr. M NOUMAN DURRANI (nouman.durrani@nu.edu.pk)

Dr. M ATIF TAHIR (atif.tahir@nu.edu.pk)

 

The increase in terrorist attacks has heightened the need for the protection of public facilities, such as airports, restaurants, and building lobbies. The identity picture on the passport is the only data which is used for identification and authentication purposes worldwide.

As a result, advancements in computer vision technology and availability of surveillance systems to autonomously detect the culprit are crucial for security.

Therefore, this project is based on real time face identification of individuals at a particular place. The flow of the project is divided into two steps:

  1. We get the image of the culprit from the installed CCTV camera video system.
  2. The video stream from the CCTV camera is broken into 2-3 seconds video and in these sets of video the provided culprit’s face is searched. This is done in parallel with other videos as well.

Furthermore, an efficient algorithm is used to detect and recognize the culprit’s face. This proposed system can be deployed at various different areas such as hospitals or offices. This system works accurately under reasonable lightning conditions and image depths.

 

 

 

 

DIGITAL TAJIR

 

BILAL SALMAN 15k-2114 (k152114@nu.edu.pk)
BILAL SHAFIQUE 15k-2152 (k152152@nu.edu.pk)
MUHAMMAD AHMED 15k-2848 (k152848@nu.edu.pk)
HAMMAD AAFTAB 15k-2873 (k152873@nu.edu.pk)

 

SUPERVISOR

Dr. RAUF AHMED SHAMS MALICK (rauf.malick@nu.edu.pk)

 

We are developing an AI based engine for personalized recommendations based on rules inferred from cognitive engine. In this project, we are fusing data from web logs, social platforms, human mobility patterns, and platform clicks along with the gestures to build a human digital persona by the development and integration of analytics pipeline for personalized effective experience for users over e-commerce platforms. Using the already existing research papers, we hypothesize that personality-based personal recommendations can be made. As products on retail websites exist in generic categories, by relabeling the products on ecommerce platforms, the product demanded by the customers can be showcased to the potential buyers and this can increase the chances of selling the products in the inventory. Similarly, by using Google Analytics, we aim to deliver a campaign management system as a personalized recommender. The landing page advertiser bridges the communication gap of the users’ demands and the product being displayed and this can increase the sales of online retailers while providing a surge in customer satisfaction.

 

SOCIAL SEARCHER

 

IBRAHIM 15k-2131 (K152131@nu.edu.pk)
TAHA ASIF 15k-2855 (K152855@nu.edu.pk)
OWAIS 15k-2190 (K152190@nu.edu.pk)
MUHAMMAD SHAHIQ QURESHI 15k-2822 (K152822@nu.edu.pk)

 

SUPERVISOR

Dr. RAUF AHMED SHAMS MALICK (rauf.malick@nu.edu.pk)

Dr. RAFI (muhammad.rafi@nu.edu.pk)

 

ABSTRACT

Social media platforms have become the common platforms where the public learns about the recent trends or news and express their emotion, voice their opinion, and provide their reaction. Put simply, social media provides the reaction to an action. The two things, the news and the response to the news are scattered between different social media platforms and news channels. A user coming to one platform will not be able to get the all the details. Different news, having a contextual link between them, are scattered and a person is not able to see the chain of events in order to get a bigger picture. Equally, their reactions are also dispersed among different social media platforms, so the general public response or different majority opinions are hidden from a person. A person is only able to see the opinions of their community. Social Searcher aims to provide a single platform which contains all the news gathered and presented in a contextual manner along with the reactions to that news. An incoming user can simply search the news and receive a complete picture of an event with all the details at a single place without spending hours in front of news channels just to clear the picture.

 

CROWDSOURCED REALTIME COMMUTE SERVICE

 

MUHAMMAD HAMZA 15k-2863 (k152863@nu.edu.pk)
MOHSIN AHMED 15k-2807 (k152807@nu.edu.pk)
HAMZA SIDDIQUI 15k-2871 (k152871@nu.edu.pk)
ANOSH 15k-2825 (k152825@nu.edu.pk)

 

SUPERVISOR

Dr. RAUF AHMED SHAMS MALICK (rauf.malick@nu.edu.pk)

 

 

This project aims to address the problem of daily commute by harnessing the potential of the concept of “crowdsourcing” through an Android application. This application will provide a platform where users can view, report, and validate all happenings and incidents related to traffic and indirectly related to their commute. Users will have the platform where they can inform and guide other users about the disturbances or breakdowns in the city’s traffic system. Also, users can validate or invalidate the updates that have been posted by other users. This will enable the users of this application to have the latest updates about the city’s traffic situation in any particular location and then plan their commute accordingly. Moreover, updates about the city’s traffic situation will also be fetched from Social Media as these sites are the hub of real-time updates and information sources from a massive user base. All this information will be available to them at one unified platform and will be just a quick glance away.

Furthermore, the basis of this project lies in leveraging the potential of mass-scale Geolocation data through the use of “Trajectory Mining”. The Android application will collect Geolocation data of the users of the application with their consent and save this data into the app’s database securely and anonymously. It will then make use of this data through mining the geolocation and trajectory of the user.

 

 

AUTISM THROUGH COMPUTING METHODS

 

MUHAMMAD GALA 15k-2875 (k152875@nu.edu.pk)
UMAR BIN ZAHID 15k-2833 (k152833@nu.edu.pk)
FAHAD NAVAID KHAN 15k-2341 (k152341@nu.edu.pk)
HAMMAD HAROON 15k-3605 (k153605@nu.edu.pk)

 

SUPERVISOR

Dr. RAUF MALIK (rauf.malik@nu.edu.pk)

 

Autism is a neurobehavioral condition in which the victim faces difficulty in having social interaction and developing language and communication skills. This behavior has a wide degree of variation in the way it affects people; therefore, it is known as Autism Spectrum Disorder (ASD). The common problems being faced by people suffering from ASD include difficulties in communication and interaction with other people. This project aims to diagnose autism at an early age which will help in initiating the treatment on timely basis. This project will be using VR technology to help in diagnosing and treating autistic people. The treatment entails improving the social interaction of the patients with various games and virtual environments.

 

 

 

 

COGIMICS – RESEARCH ON COGNITIVE ENGINE USING VISUAL, TEXTUAL AND PHYSICAL EXPRESSIONS

 

SAJJAL MALIK 16K-3983 (K163983@nu.edu.pk)
IBTISAM AHMED QURESHI 15K-2187 (K152187@nu.edu.pk)
ZAKI USMANI 15K-2887 (K152887@nu.edu.pk)

 

 

SUPERVISOR

Dr. RAUF SHAMS MALICK (rauf.malick@nu.edu.pk)

 

Today, all major digital brands struggle to understand their customers. Physically, salesmanship perceives the human cues and adapts accordingly which is the greatest lag faced by online retailers. Digital body signatures in the form of gestures and swipes are used to make emotions and personality predictions to develop a cognitive engine. To bridge the gap between the digital brands and their customers, the technology will enhance the customer experience by predicting the user’s personality and then developing personalized recommendation systems. This cognitive engine shall comprise of emotion prediction, gender prediction, and personality prediction to get a holistic viewpoint of the user preferences and choices. This cognitive engine shall get a holistic viewpoint of the user preferences and choices. A dataset based on images, swipes, and social text were collected from 100 people along with their BFI-10/BFI forms. The prediction model was tested on the dataset which was validated against their original personalities identified by the BFI forms. This service will be a component that can be further integrated into digital marketing, ecommerce platforms, and immersive gaming.

 

 

 

CRICKETER’S TALENT HUNT: A SMART RECOMMENDATION SYSTEM

 

MUHAMMAD SHEHROZ 15k-2249    (k152249@nu.edu.pk)
SAQIB AHMED KHAN     15k-2810                             (k152810@nu.edu.pk)
                            BILAL YAQOOB               15k-2207   (k152207@nu.edu.pk)

 

SUPERVISOR

Dr. FAHAD SAMAD(fahad.samad@nu.edu.pk)

Today’s traditional cricket selection system on International, Domestic, and Club level is based on an approach which is manual, and can be biased, inefficient and considers very limited statistics of players. Our system consists of a mobile application which provides opportunities to emerging players to present their talent. Furthermore, it reduces burden of the trial system for the cricket representatives to select cricket players. These representatives will include Pakistan Cricket Board Representatives, Regional Representatives, and Club Representatives. With the use of this system, the player will create his profile and will feed his statistics, which will be verified by social media module which includes likes or comments on his profile or his posts to verify the player’s statistics. Furthermore, each and every representative will get top recommended players based on their verified performance and career statistics. They can also communicate with each other through a chat module.

 

 

 

 

 

SMART RECRUITANT

 

SHAHZAIB IBRAHIM 15k-2272 (K152272@nu.edu.pk)
MUNEEB UR REHMAN 15k-2283 (K152283@nu.edu.pk)
MUDASSIR ASHRAF 15k-2834 (K152834@nu.edu.pk)
DANIYAL AHMED 15k-2893 (K152893@nu.edu.pk)

 

SUPERVISOR

Dr. FAHAD SAMAD (fahad.samad@nu.edu.pk)

 

Conducting interview of every interested candidate for job is a real problem for the interviewer. It is costly and takes a lot of time and resources to conduct the hiring procedure. The main purpose of our project is to conduct interviews and judge candidates based on their credentials/skills and the job requirement. The interview will be conducted by the bot, evaluating both technical and non-technical skills. The results will be generated by the system and will rank the candidates accordingly. Recruiters can call the top candidates for the final interview. Our system will have two interfaces:

1.             Job thread:

Where the employer will post the job and the candidates will apply for it.

2.             Conversational interface:

For conducting interview of the candidate.

This software will serve companies in the recruitment process. To use the Recruitant service, companies will be required to pay a fee.

 

INTEGRATED SERVICES PROVIDING SYSTEM

 

              AHMED HUSSAIN 15k-2149 (k152149@nu.edu.pk)
              ALI TAHIR15k-2181 (k152181@nu.edu.pk)
              AMMAR RASHID SIDDIQUI15k-2288 (k152288@nu.edu.pk)
              SYED MAZHAR ABBAS15k-2252 (k152252@nu.edu.pk)

 

SUPERVISOR

Dr. FAHAD SAMAD (fahad.samad@nu.edu.pk)

Mr. JAZIB UL HASSAN

 

 

Smart Phones have become extremely powerful in the recent years. Dramatics breakthroughs in processing power along with number of extra features included in these devices have opened the doors to untangle a wide range of uneventful problems. One application that falls into this category is the Pandora, which not only helps to resolve the problems but also provides an amenity facility of Handyman at your door steps. The main purpose of Pandora is to report the problems faced by the residents of the city and getting reply from the District Municipal Corporation. This is the proper medium of communication between the people and the government through which citizens could put forward their requests of fixing the problems. The user has the facility to report different types of problems, moreover he/she can view, like, dislike and comment the problems reported by others in his area. The other purpose of Pandora is to facilitate the user with the different types of handymen, such as Carpenter, Plumber, Electrician in an optimized way. The additional feature allows the user to place the handymen on shelf for a time convenient for him.

 

 

SCHOOL RECOMMENDER

 

ALI ARSLAN 15k-2113 (K152113@nu.edu.pk)
FAISAL GULZAR 15k-2338 (K152338@nu.edu.pk)
ROHAN RASHID 15k-2348 (K152348@nu.edu.pk)
SAIF ALI HAZRAT 15k-2340 (K152340@nu.edu.pk)

 

 

 

SUPERVISOR

Dr. FAHAD SAMAD (fahad.samad@nu.edu.pk)

 

 

In a country like Pakistan where process of finding and selecting school is manual, parents often get exhausted. Unfortunately, the only medium they have in finding that perfect school is by getting opinions from family and friends. Most of the time after completing admission formalities, students and parents realize that the selected school doesn’t meet their requirements. This gave us the motivation for developing this project. For solving this problem and to assist parents, we developed a web based school recommender system that suggests a perfect school to the user. A survey was conducted from parents and their needs and requirements of school were asked. Those needs were included as priorities. Information of school located in areas of Karachi such as gulshan-e-Iqbal, Defence and north nazimabad were collected. The collected information includes fees, location, religious affiliation etc. Those Information were entered in the system. The recommendation of school was then done on the priorities given by parents. Parents were able to provide weightage to different priorities such as fees, location, education system. School that matches the user priory is suggested and an option for submitting admission request is also provided. The result of the recommendation depends on the criteria of parents so recommendation will differ for everyone. In future more features can be added in the system such as appointment booking for admission interview or submission of documents to the school directly through the system etc.

 

 

 

FYP WORK FLOW ENGINE

 

MUHMMAD BURHAN ZAHID 14k-2046 K142046@nu.edu.pk
FAIZAN AHMED QUADIR 14k-2852 K142852@nu.edu.pk
SHEHROZ HAMEED 14k-2318 K142318@nu.edu.pk
RAKESH KUMAR 14k-2858 K142858@nu.edu.pk

 

 

SUPERVISOR

Dr. ABDUL AZIZ (abdulaziz@nu.edu.pk)

Mr. SHAHBAZ AKHTAR SIDDIQUI (Shahbaz.siddiqui@nu.edu.pk)

 

The world is shifting from file-based system to database and shifting from paper environment to paperless environment. Yet some processes still require human interaction and too much paper. Here, we are addressing one process which consumes both paper and time. The ultimate goal of this project is to make this process paperless and minimize the human interaction. The project is a web-based solution which consists of different modules such as the student module, the jury module, the Supervisor module and the coordinator module. Through this project student can register himself for FYP, select supervisor, select group members, can check their marks, and can also compare his project idea with others. Similarly, jury can upload marks. Coordinator can send announcement and related document templates. Supervisor can upload marks, check documents as well as send announcement to students.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

INTERIOR DESIGN IN AR

 

 

SHAYAN ABID 15k-2150 (K152150@nu.edu.pk)
ABDULLAH WASEEM 15k-2177 (K152177@nu.edu.pk)
SARA UROOJ 15k-2232 (K152232@nu.edu.pk)

 

 

SUPERVISOR

ENGR. ABDUL RAHMAN MAHMOOD (ABDULRAHMAN@NU.EDU.PK)

 

 

Augmented reality is a technology that combines virtual reality with the real world in the form of live video imagery that is digitally enhanced with computer-generated graphics. AR can be experienced through headsets that people wear and through displays of mobile devices. The number of AR based application both for iPhone and Android devices has expanded dramatically, and they aren’t limited to games. Retailers are showing tremendous interest in AR possibilities. Apart from games and fun purposes, AR is being widely used in military and medical sciences. It is also being used for navigation, where the smartphone’s camera in combination with the GPS, shows the selected route over the live view of what is in front of the car, and for sightseeing in tourism industry.

Interior Design App is also an AR based application which will map any of the user desired product in the environment, will show the user virtually that how their place will look after all renovation and decoration, and the best part is that user will also be able to buy all of the products which they love for their space.

The purpose of working on such application is to overcome the defects which were in previously made apps and to safe the users from all the fuss they have to deal with while renovating their space.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EZ COMPLAINT – COMMUNITY

ENGAGEMENT PORTAL

 

MUKANDER KUMAR 15k-2198 (K152198@nu.edu.pk)
NIRMAL 15k-2210 (K152210@nu.edu.pk)
SHAKTI 15k-2898 (K152898@nu.edu.pk)

 

 

SUPERVISOR

ENGR. ABDUL RAHMAN MAHMOOD (abdulrahman@nu.edu.pk)

 

In current time period where everyone has a very busy life schedule, every day we see certain situations where people face problems in many departments like education, gas, electricity, social affairs. They want to register complaints against these problems but none of us goes to these departments and register complaint because everyone is engaged in his/her own busy routine. It made us think that why not make such a platform where people can register their complaints without compromising their busy schedule by going to these departments and submitting applications. Our vision is to link all government sector departments under a unique platform so that community can register its complaints according to that particular department by just recording an audio clip in roman Urdu. Because our target users are the people of Pakistan, Roman Urdu audio clip will be easy for all the community to type its complaints to save time. When user register the complaint, his/her complaint is immediately transferred to that particular department and user will be notified about their complaint status.

 

 

 

 

VR SKILL SET FOR INTELLECTUAL DISABLE PEOPLE

 

SHAIKH MUHAMMAD UMAIR 15k-2884 (K152884@nu.edu.pk)
MUHAMMAD RAFAY  15k-2289 (K152289@nu.edu.pk)
AHSAN ZAID 15k-2829 (K152829@nu.edu.pk)

 

SUPERVISOR

ENGR. ABDUL RAHMAN MAHMOOD (abdulrahman@nu.edu.pk)

 

Virtual Reality is a technology that combines both ‘virtual’ and ‘reality’ world together. Virtual is defined by its name that means near and reality is referred to as what we experience as human beings. The term Virtual Reality means NEAR-REALITY. In VR we experience computer-generated simulation of a three-dimensional environment that can be experienced by user with the help of headset that people wear with the help of mobile device attached to the headset and with the display on mobile devices. Some of the VR devices can be:

  • Google Card Board
  • Samsung Gear VR
  • Oculus

The number of VR applications are increasing day by day but still its development is quite slow as compare to Android/IOS app. Since it’s a new technology so the development in VR is a bit slow but still a number of VR applications are being made available. The reason is not everyone is familiar with the development of this technology that how to map real world into virtual world, what tools to use and how to make things visible, available, and how to make things that experience user the same experience

As he is experience in the real world. That how it feels to be in the environment and feel things around.

The number of VR application are based on both Android Application and IOS Application in addition to that it’s also made for Oculus because it is a headset which doesn’t allow any mobile devices to be inserted in that. A part from game development in VR, VR is also used for people to train some skill set which may result in increase in productive with respect to time and multiple resources.

VR Skill Set for Intellectual Disable is also an VR based application which will map the real world into virtual world which show the user things in the virtual environment and will be able to learn skill set and able to map things in reality. They can achieve training in their desired skill set and the Intellectual Disable person will be able to learn skill set with minimum resources and without the loss of life and saving materials according to their desired skill set. Few resources can be:

  • Time
  • Money
  • Materials
  • Equipment
  • Trainings

The purpose of working on such application is to make the Intellectual Disable person is to learn skill set and life a reputable life in the society and to earn something and live their life easily.

 

 

 

 

STUDENT HEALTH EDUCATION (SHE)

 

AQUIB ALI AHMED 15k-2343 (K152343@nu.edu.pk)
                                       BATOOL SALEEM 15k-3627 (K153627@nu.edu.pk)
ANMOL AGHA 15k-2850 (K152850@nu.edu.pk)
FATIMA RASHEED 15k-2295 (K152295@nu.edu.pk)

 

SUPERVISOR

ENGR. ABDUL RAHMAN (abdulrahman@nu.edu.pk)

 

Our project’s aim is about the health awareness education for women where we guide them about H2S (Health, Hygiene, Sanitation). We will provide them a platform where they can discuss their health problems with the lady doctor or gynecologist and get prescription from them for free and the pharmacy will also be linked with us which will provide medicine to the doorstep to those women if they want to on COD (cash on delivery) basis. The women talk about their health problems with doctor by voice note or by messaging option and can get prescription from the doctor.

The second part of our project is for primary school children we will make software where the health report will be maintaining for each student where the health survey team on monthly basis will update it after checkup. Students and the school authority can view the health report of the student but it can only be updated by the HST (Health survey team).

 

 

 

 

SECURE COMMUNICATION FOR LOW POWERED IOT DEVICES IN NARROWBAND ENVIRONMENT

 

ABDUL MUQEET SHAHID 15k-3607 (k153607@nu.edu.pk)
MUHAMMAD USMAN KHOKHAR 15k-2344 (k152344@nu.edu.pk)
HIRA FATIMA 15k-3621 (k153621@nu.edu.pk)
HAMMAD MASOOD 15k-2885 (k152885@nu.edu.pk)

 

SUPERVISOR

ENGR. ABDUL REHMAN (abdulrahman@nu.edu.pk)

There are many emerging areas in which highly constrained devices are interconnected and communicated to accomplish some tasks. Narrowband Internet of Things (Nb IoT) enables many low resources and constrained devices to communicate, compute process and make a decision in the communication network recently. Nb IoT devices are rapidly becoming ubiquitous while IoT services are becoming pervasive. Their success has not gone unnoticed and the number of threats and attacks against Nb IoT devices and services are on the increase as well. Cyber-attacks are not new to Nb IoT, but as Nb IoT will be deeply interwoven in our lives and societies, it is becoming necessary to step up and take cyber defence seriously. Hence, there is a real need to secure IoT. In the heterogeneous environments for IoT, there are many challenges and issues like power consumption of devices, limited battery, memory space, performance cost, and mainly the security in the Information Communication Technology (ICT) network. In this attempt, we will illustrate a state-of-art of lightweight cryptographic primitives which include lightweight block cyphers, hash function, stream cyphers, high-performance system, and low resources device for IoT environment in details. We analyze many lightweight cryptographic algorithms based on their key size, block size, number of rounds, and structures. In addition, we discuss the security architecture in Nb IoT for constrained devices environment and focus on research challenges, issues and solutions. Finally, a proposed security scheme with a service scenario for an improvement of resource-constrained IoT environment and open issues are discussed.

 

 

 

 

KEYNOTE SUPERIMPOSITION AND VISUAL FEEDBACK BASED PIANO LEARNING SYSTEM USING AUGMENTED REALITY

 

UZAIR KHAN 15K-2183 (K152183@nu.edu.pk)
ABDUL QADIR 15K-2182 (K152182@nu.edu.pk)
SHUJA ABBAS 15K-2303 (K152303@nu.edu.pk)
SHIZRA AIJAZ 15K-2869 (K152869@nu.edu.pk)

 

SUPERVISOR

Mr. FARRUKH HASAN SYED (farrukh.hassan@nu.edu.pk)

Mr. SAYED YOUSUF (sayedyousuf@yahoo.com)

 

Piano learning resources are widely available over the internet. Currently, piano learners who are not advanced enough to play without looking at the keys must simultaneously look at an external piano sheet / online video resource while mentally translating the required chords to play on their instrument. There exists no complete solution that provides a user with an immersive learning environment with music feedback that does not require the user to divert their attention between the musical instrument and the training resource. We propose an AR based piano learning system that provides visual feedback in the form of keynote superimposition for what keys to play next, as well as the accuracy of the keys that the learner just played. The target audience is beginner piano learners.

 

 

 

AR-GO

 

RABI UZ ZAMAN 15k-2315 (k152315@nu.edu.pk)
OSAMA MAZHAR 15k-2278 (k152278@nu.edu.pk)
WAJAHAT KHAN 15k-2811 (k152811@nu.edu.pk)
 AFFAN AHMED 15k-2161 (k152161@nu.edu.pk)

 

SUPERVISOR

Mr. FARRUKH HASAN SYED (farrukh.hassan@nu.edu.pk)

Mr. SAYED YOUSAF (sayed.yousaf@nu.edu.pk)

 

Augmented reality is taking the market by fire as many companies have started building their own applications related to augmented reality. Different businesses are also showing interest in AR applications. Superstores are using augmented reality in their stores to enhance the shopping experience of their customers. Similarly, our application provides companies a way of promoting new products using augmented reality.

 

Our project aims to introduce a gaming like experience to help our user to bring some relief in their hectic life and add some novel experience into their shopping. Multiple stores can use our platform and can update their information themselves.

 

 

 

ADEQUATE BODY SUITING EXPERIENCE

 

SYED MUHAMMAD HUMRAAN 15k-2882 (k152882@nu.edu.pk)
ALEEM 15k-2847 (k152847@nu.edu.pk)
AZEEM KHALID 15k-2845 (k152845@nu.edu.pk)

 

SUPERVISOR

Mr. FARRUKH HASAN SYED (farrukh.hassan@nu.edu.pk)

Mr. SAYED YOUSUF  (sayedyousuf@nu.edu.pk)

 

Adequate Body Suiting is a virtual dressing room experience typically designed to replace the old-fashioned shopping mall dressing rooms with a newly developed and advanced technology based on augmented reality in which a person can easily visualize various types of wardrobes available in a particular shop closet, just by using hand gestures. This dressing room will first figure out the dimensions of a person standing in front of Kinect and then it will show the customer all the collection of wardrobe available according to their particular body size. Our wardrobe closet will include caps, glasses, tops, bottoms and shoes. In future this Adequate Body Suiting technology will completely wipe out the concept of a dressing room from shopping arenas with their innovative virtual dressing room experience.

 

 

 

 

 

INVESTIGATION OF MULTI-TARGET LEARNING

 

MIR DANISH SHAHAB 15k-2891 (k152891@nu.edu.pk)
AHMED SAJJAD 15k-2872 (k152872@nu.edu.pk)
SYED FAROOQ AHMED 15k-2166 (k152166@nu.edu.pk)

 

SUPERVISOR

Mr. FARRUKH HASAN SYED (farrukh.hassan@nu.edu.pk)

 

The goal of this research is to investigate multi target learning. In multi target learning we have more than one targets to predict. Specifically, this research will investigate relations between different targets, how targets affect each other and how their interactions affect the results. The second investigation is related to the situation when there are a different number of labelled examples available for different targets and whether applying different multi-target algorithms to different targets would yield better results.

 

 

 

FOODSQUARE

 

ABDULLAH 15k-2129 (k152129@nu.edu.pk)
SAKINA ABBAS 15k-2281 (k152281@nu.edu.pk)
M. UMAR MOIN 15k-2802 (k152802@nu.edu.pk)

 

SUPERVISOR

Mr. MUHAMMAD SHAHZAD (mshahzad@nu.edu.pk)

Dr. RAUF MALICK (rauf.malick@nu.edu.pk)

 

The online food industry is one of the fastest growing industries. While the most fundamental function still remains online food ordering from food vendors, the real question is as to what should ease a user in ordering food online. The major problems a typical user faces while using a food ordering application are a) searching a restaurant, and b) searching a particular food item. Upon searching for a restaurant, either by name or by location, the search results inflating the screen are haphazardly listed, thereby lacking a ranking mechanism to priorities one search result over the other. Moreover, due to an absence of a search mechanism for a particular food item, users often find it difficult to compare a food item, e.g. Burger between the different vendors serving it. FoodSquare, an online food search engine, aims to mitigate these problems by utilizing user reviews, employing analytics and identifying user personas to rank a user’s search results and personalize them according to his/her preferences by also providing them recommendations based on association rule mining.

 

 

 

 

DIABETIC RETINOPATHY (DR) CLASSIFICATION THROUGH SMARTPHONE FUNDUS IMAGING

 

IBRAHIM CHIPPA  15k-2329 (K152329@nu.edu.pk)
ASAD ANWAR  15k-2325 (K152325@nu.edu.pk)
MARIUM SALMAN  15k-2411 (K152411@nu.edu.pk)
UMER ABDULLAH 15k-2350 (K152350@nu.edu.pk)

 

SUPERVISOR

MUHAMAD SHAHZAD (mshahzad@nu.edu.pk)

  1. RAUF SHAMS MALICK (rauf.malick@nu.edu.pk)

 

Each year about 12% of blindness in the US is due to diabetic retinopathy, with numbers set to rise by the year 2030. It is among the leading cause of blindness among patients aged 20-64 years who have been suffering from diabetes. Among 90% of cases can be reduced with timely diagnoses which is through fundus imaging. In Pakistan many cases of DR go unnoticed and eventually untreated, due to un-availability of equipment required for fundus screening; which is the initial step in diagnoses and treatment of DR. Through research it was observed that smartphones-being affordable and portable- can be a possible alternative to the expensive machinery required for DR diagnosis. Hence, by integrating image processing and deep learning techniques an approach has been taken allowing smartphones to diagnose the possible presence of DR in diabetic patients. Multiple CNN architectures have been trained and tested on the Kaggle dataset for the diabetic retinopathy contest.  Among the architectures trained and validated Resnet50 provided an accuracy of 74%.

 

DATA ANALYTICS AND CLASSIFICATION FOR ‘HYSAB KYTAB’

 

RAJNESH KUMAR 15k-2133 (k152133@nu.edu.pk)
AJAY KARERA 15k-2229 (k152229.edu.pk)
KHANZADA BEHJAT HUMA 15k-2379 (k152379nu.edu.pk)

 

SUPERVISOR

Mr. MUHAMMAD SHAHZAD (mshahzad@nu.edu.pk)

Mr. ABDUL HASEEB (abdul.haseeb@jaffer.com)

 

The application ‘Hysab Kytab’ is a personal finance management system which helps its users record income and expense transaction to keep track of their financial statements and achieve their saving goals. Each Transaction record has several features such as description, tag, category, and account which can be selected either from the predefined categories or the user can make a custom category using free text. For example, a user enters Meezan bank’s in account feature which actually is ‘Meezan Bank’. Therefore, the use of custom text and roman Urdu in certain features makes it difficult to analyze the system’s huge free text data to further improvise the application. For analysis, predefined pattern matching with the custom text, substring matching using natural language processing techniques and similarity index of the free text with predefined classes helped to understand the statistical aspects of the given dataset. To encounter this problem, a classification model using the statistical analysis as the building blocks maps the customized classes of categories to predefined categories which serve the business purposes of the application.

 

 

 

ATTENDANCE SYSTEM USING CCTV CAMERAS

 

ABDUL MUQSIT ABBASI  15K-2279 (K152183@nu.edu.pk)
MANAL FAROOQ  15K-2849 (K152182@nu.edu.pk)
JUNAID MASROOR  15K-2228 (K152303@nu.edu.pk)
TAHA YAHYA BAIG  15K-2877 (K152869@nu.edu.pk)

 

 

 

SUPERVISOR:

Mr. MUHAMMAD SHAHZAD (mshahzad@nu.edu.pk)

 

Dr. ATIF TAHIR (atif.tahir@nu.edu.pk)

 

Facial Detection and Recognition performed on Real-time CCTV Footage to mark the attendance of employees, is a deep learning project that uses specific face detection and recognition algorithms to identify and recognize the employees of an organization as they enter or exit the building and mark their attendance accordingly. Currently, the attendance system in FAST-NUCES is a manual system that is time consuming and error prone where an employee has to mark their own attendance in a specially maintained register at least twice; once entering the university and second while leaving. Live CCTV footage is fed into a specifically designed deep learning (DL) model to first detect and then identify the employees as they enter and leave the university premises. Finally, it marks the attendance of the detected personnel and alert them via an Email. The desktop application is written in Python.

 

 

 

 

 

PERSONALIZED TREATMENT OF CANCER USING MACHINE LEARNING

 

OSAMA ARIF 15k-2892 (K152892@nu.edu.pk)
SAAD AHMAD 15k-2881 (K152881@nu.edu.pk)
FAIZAN RASOOL 15k-2312 (K152312@nu.edu.pk)
ZAIN ALI KHAN 15k-2299 (K152299@nu.edu.pk)

 

SUPERVISOR

Mr. MUHAMMAD SHAHZAD (mshahzad@nu.edu.pk)

Dr. RAUF SHAMS MALICK (rauf.malick@nu.edu.pk)

 

Diseases that appear pathologically similar often responds differently to same drugs. Conventionally, doctors use clinical experience to prescribe drugs to patients i.e. “one shoe fits all”. In some cases of cancer, doctors use hit-and-trial method to identify the most suitable drug for the patient, hence, wasting critical time. Precision or personalized medicine rejects this idea of one drug for all patients, and instead it separates patients into different categories, these categories are based on the amount of resistance a patient’s disease exhibits and the risk of their disease. Primary goal of precision medicine is to apply accurate diagnosing technique based on the patient’s available genetic information and its behavior to the drugs available for diagnosis. We intend to demonstrate a promising model to identify the sturdy bio-markers and predict optimal drugs for personalized treatment of particular cancer using public data-sets from cancer patients, and genomic and drug databases.

Since the conventional treatments are mostly arbitrary and non-parametric, we cannot determine the actual accuracy of the conventional treatments. On the contrary, by taking into account huge data sets precision medicine models are probabilistic and provide much higher accuracy. Moreover, once personalized medicine methods are deployed, the amount of severe disease which aren’t treated successfully can reduce exponentially. Cancer, which is an unpredictable disease can also be diagnosed, categorized, and treated at a very early stage.

 

 

 

 

 

 

 

APP OF KINDNESS

 

NOMAN 14k-2289 (K142289@nu.edu.pk)
M. UMAR SALEEM 13k-2343 (K132343@nu.edu.pk)
M. ALI AHMED 14k-2838 (K142838@nu.edu.pk)
ALLAWDIN 14k-2824 (K142824@nu.edu.pk)

 

SUPERVISOR

Mr. SHAHBAZ SIDDIQUI (shahbaz.siddiqui@nu.edu.pk)

 

 

App of Kindness will provide a platform to users to provide and accept donations to an individual or an institution or even a charity-based organization. As a donor, they will choose to donate an item which no longer is in their use to a person in need by either choosing from the nearby map location for any acceptor or by browsing through a sum of descriptions in a categorized item list. As an acceptor, they can either put up a description requesting for the donation of the required item or similar to the donor, look for a donor in the nearby-area. The app’s purpose is to bring the donor closer to the acceptor using modern day computer science based applications as to facilitate consumers with need inspired by the Wall of Kindness, where any person could leave items ranged from clothes to food against a wall for people in need. Our project is a spin-off to this idea and is for the same people who are in need of essentials that they cannot afford. We will allow the user to put up a description of the items they are willing to donate or are looking for to accept as a donation. This will result in the direct communication between the donor and acceptor and depending on the donor, they get to choose where, to whom and why to donate the item to. Further development on this platform will require collaboration with qualitative institutes based on helping people like charities or schools etc.

 

 

 

 

SMART BIKE

 

 

                             MUSAWWIR MAQSOOD 15k-2318 k152318@nu.edu.pk
                             M.MUJTABA QUAIYUM 15k-2147 k152147@nu.edu.pk
                                  MUHAMMAD OWAIS 15k-2118 k152118@nu.edu.pk

 

 

SUPERVISOR

 

Mr. SHAHBAZ SIDDIQUI (shahbaz.siddiqui@nu.edu.pk)

 

 

The idea of this project is to provide ease and safety to bike riders by providing basic details of the bike in a simple and user-friendly interface. The basic details of the bike include bike’s tire pressure, petrol level indication and headlights condition which will be displayed on the dashboard. This project will also help the rider while riding the bike by providing the details of the routes and the destinations. The user will also be prompted with an alert if the petrol level drops down the emergency level and, it will locate the nearest petrol pump and display on the dashboard. Another feature of this project is bike pooling system. The biker i.e. the one who owns a bike can enter his/her destination and enables the ride sharing option and the one who is willing to get a ride can accept his/her share and thus he/she will be picked up by the rider as location of both the rider and the customer will be shared to each other.

 

 

 

UNIVERSITY SMART CARD SYSTEM

 

ALI HAIDER 15k-2347 (k152347@nu.edu.pk)
ZAIN UD DIN 15k-2812 (k152812@nu.edu.pk)
MUHAMMAD UMER 15k-2132 (k152132@nu.edu.pk)
SAAD ALAM 14k-2004 (k142004@nu.edu.pk)

 

SUPERVISOR

Mr. SHAHBAZ AKHTAR SIDDIQUI (shahbaz.siddiqui@nu.edu.pk)

 

Nowadays, students and teachers at FAST university tend to carry their university identity card, automatic machine teller cards (ATM), money and documents while they are on the premises. If a person needs to buy some food from the cafeteria, he needs to have change amount and sometimes he doesn’t have a change or the exact amount to be paid. In the same way, a person needs to have some change money in order to pay his fines and dues. In other words, a person needs to have all of these important things with him at all times.

 

Each individual must present his identity card at the time of entering the premises which are not completely authenticated and even an unknown person can use a fake identity card to enter the campus. Moreover, teachers manually need to mark the attendance of each class which takes at least 5 to 10 minutes. What if the attendance of the entire class can be marked in less than 5 minutes? What if there is no need to carry out cash and important documents? What if there is no need to stand in queues to get your order in the cafeteria? And what if the librarian doesn’t need to manually search for your record?

 

This smart card provides an IoT based system that can be controlled by the administration and every individual of FAST University for various purposes as discussed above. The main idea behind this project is to provide the students and teachers of FAST University with a more secure and multipurpose smart card that can be used within the premises without the need to carry a bundle of items discussed above.

The whole system will consist of a Smart identification card required by each student to enter the premises and access the functionalities of the card. Both administration and parents will have the privilege to check the status of their smart card through an android application.

 

 

 

 

JOB ASSISTANCE APPLICATION

 

SYED HAMZA LIAQUAT 15k-3628 K153628@nu.edu.pk
MUHAMMAD HASSAN QURESHI 15k-2522 K152522@nu.edu.pk
                                       OSAMA SHEIKH   13K-2199 K132199@nu.edu.pk
HAMMAD VOHRA 14k-2267 K142267@nu.edu.pk

 

SUPERVISOR

MR SHAHBAZ SIDDIQUI       Shahbaz.siddiqui@nu.edu.pk

 

 

The issue of hiring skilled workers related to different walks of life like construction labors and mechanics is an emerging problem. Unemployment is increasing at a continuous rate although we have skilled man power available in each area. This project is designed to transform how a person finds works and earn money through reliable ways, and provide the ease to user without hesitation of going in local market and search the related person. This project can bridge the gap between users and workers.

This project will provide many job opportunities to daily wages workers. It can also provide part time job to many people. In addition, it can also save workers’ time and cost by providing pin point location of their working site at hand held device i.e. mobile phones. Employer will be benefited too as they not need to search worker at different places. User can call worker according to their service type. Apart from this, the project will also help Contractors to hire labor that is required at various situations like when they have shortage of workers

Our project is for mobile users. In this application, there are two modules. The first one is for users like contractors / while second module is for labours. Both modules being maintained by admin of this application. The front end of the application is based on Android studios version 3.1.2, and in Back end node server is working with firebase database. New features can be added in this application according to the needs of users. We make separate IOS application for the apple users to facilitate the ease and to broaden the scope of our project.

 

 

 

WING DING (A Get-Together Application)

 

SYED ABRAR AHMED  14k-2275 (k142275@nu.edu.pk)
MUHAMMAD TALIB14k-2167 (k142167@nu.edu.pk)
SYED WASAY SAEED14k-2146 (k142146@nu.edu.pk)
OSAMA NASIR  14k-2248 (k142248@nu.edu.pk)

 

SUPERVISOR

Mr. UBAID AFTAB CHAWLA (ubaid.aftab@nu.edu.pk)

Mr. SAYED YOUSUF         (sayed.yousuf@nu.edu.pk)

 

Wing Ding (A Get-Together Application) is a one of a kind social platform that is based on connecting people in every aspect that is related to an outing, social event, hosting a party or even fine dining. The main aim of this application is to help provide maximum comfort and ease to the users to reach their particular destinations as well as covering other aspects for the success factor of an event. This project not only covers the tracking location functionality of users, it also provides them the option to interact with each other, reduces the hassle of providing discomfort to anyone regarding their pictures, keeps them updated of social events near them that they might be interested in as well as have an estimation feature and reservation management system for fine dining etc. In order to do this there are many factors that must be taken into accounts especially the factor of a user having an android cell and being adept at using it at a moderate level, (adept; being able to use other functionalities of a smart phone except calling and texting and being familiar with applications). This application is an important asset to the current generation who are deeply immersed in attending social activities whether it is for an academic reason or for entertainment purposes

 

 

 

 

”KAIZEN”-CARPOOL APPLICATION

 

TARIQ PARVEZ 15k-2171 (k152171@nu.edu.pk)
DANIYAL TAHIR 15k-2271 (k152271@nu.edu.pk)
UMAR ARSHAD 15k-2801 (k152801@nu.edu.pk)
M.HUZAIFA SIDDIQUI 15k-2201 (k152201@nu.edu.pk)

 

SUPERVISOR

Mr. UBAID AFTAB CHAWLA (ubaid.aftab@nu.edu.pk)

Mr. SAYED YOUSUF (sayedyousuf@yahoo.com)

 

Ride-sharing is an important component of sustainable urban transportation as it increases vehicle utilization while reducing road congestion. By sharing rides, driver’s offer available seats in their vehicles to passengers who want to travel in similar directions or localities. Ride sharing helps users save money, reduce traffic congestion, reduce our carbon footprint and promote the benefits of social networking. Traditional ride-sharing approaches are suitable for long-distance travel, especially inter-city travel, yet they are not flexible enough for short routes within cities. Carpooling is an easy and effective way to reduce the above mentioned problems. Depending on the size of carpool and distance of one’s commute, individuals could save thousands of rupees every year. But carpooling does not just benefit the wallet; it can positively affect the environment and the traffic in urban areas. Further advantages include:

 

  • Reduce your carbon footprint: According to transport environment.org; every car annually emits its own weight in carbon dioxide into the atmosphere. If people carpool to and from work, they could probably cut the transportation emissions number in half, or even better.

 

  • Usable commute time: Driving demands our complete attention, which means its difficult to be productive with your time on the road. Riding in a carpool allows you to check and answer email, read the news, check bank balances, or even close your eyes for a few minutes before starting your day.

 

  • Save Money: Individuals will save by using less gas. By carpooling as using less gas, you will be able to spend your money on other things, pay off your debts, and increase your savings.

 

  • Reduce traffic congestion: The benefits of carpooling on a large scale are huge. Every carpooling participant takes another car off the road, which means less congested roads and highways. As more people take part in carpooling programs or organize carpools within their workplaces, overall traffic congestion will decrease, which reduces fuel consumption, commute time, and the cost of road repairs.

 

 

Our goal is to develop a service that enables dynamic ride-sharing as part of sustainable urban mobility. Dynamic ride-sharing refers to a service that matches ride requests and ride offers on short notice with/without prior agreement between driver and passenger. Our ride-sharing system aims to bring together travelers with similar itineraries and time schedules on short-notice.

 

 

 

INTELLIGENT – AUTOMATED WAITER SYSTEM

 

BILAL ARIF15K-2184 (k152184@nu.edu.pk)
RIAZ JAFFARY 15K-2155 (k152155@nu.edu.pk)
UMAIR KHAN 15K-2128 (k152128@nu.edu.pk)
 DAWAR KAZI15K-2895 (k152895@nu.edu.pk)

 

SUPERVISOR

Mr. UBAID AFTAB CHAWLA(ubaid.aftab@gmail.com)


The systems currently being used in restaurants have only a simple POS system, which has no customer interaction. These systems are mostly dependent on waiters, which is why a large number of waiters are required, in order to accommodate all the customers efficiently and effectively. The Intelligent Automated Waiter System is more than a traditional POS system. This system will provide real time user interaction and customer assistance. It consists of an android application for customers and a desktop based POS and management system for the restaurant managers and cashiers. The customers will use the application to reserve the table, view the menu, place orders, calls the waiter for assistance, and call for bill. This system will help manage and run the restaurant systematically and reduce human errors and increase accuracy. Additionally, number of waiters required in the restaurant will also be decreased, which will save the labor cost.

 

 

 

3D DRAWING IN AUGMENTED REALITY

 

M.ABDULLAH 15k-2255 k152255@nu.edu.pk
AFAQ JAVED15k-2196 k152196@nu.edu.pk
HAMMAD JAFFRI 15k-2194 k152194@nu.edu.pk
MUSTAFA HUSSAIN 15k-2108 k152108@nu.edu.pk

 

SUPERVISOR

Mr. UBAID CHAWLA ubaid.aftab@nu.edu.pk

Mr. SAYED YOSUF     sayed.yosuf@nu.edu.pk

 

3D structuring in augmented reality is an android application which will allow users to check and test 3D physical structures like: table, box, or other rigid physical shapes in their real 3D environment having 360 degrees calibration. It will follow actual world physics laws mainly: light sensing, gravity, magnetic calibration, acceleration, orientation in 3D space. Users will be able to pre-check things before setting them. The main feature is that users will be able to draw objects manually using tools like in MS-Paint, 3D max or other software. Tools will be implemented using OpenGL Core library, and that objects will be checked in real-environment using AR technology. This application will revolutionize market in terms of buying and selling useful objects. 3D architects will be able to draw project structures and check them instantly. Hence, this application would play an important part in shopping and planning.

 

SMART SHOP

 

OWAIS KALIMI                               15k-1234 (k152304@nu.edu.pk)
   SYED HASAN HAIDER                  15k-1058 (k152282@nu.edu.pk)
MUHAMMAD OWAIS FIRDOUSI 15k-1016 (k152225@nu.edu.pk)
SYED SHAYAN HAIDER                15k-1341 (k142150@nu.edu.pk)

 

 

 

SUPERVISOR

UBAID AFTAB CHAWLA (ubaid.aftab@nu.edu.pk)

 

 

In the modern world, people are more attracted towards the online market place to save their time from the hassle to search their required products at the super stores. This shift in the dynamics of the market place raises the risk of the downfall of the share of physical existence of the super markets. To stabilize and support the conventional super markets, this project facilitates the customers as well as the management to minimize the waiting period of the customers at the billing counters and to guide them the location of their desired products within the superstore.

 

The customer has to install the customer application by which he can scan the products he wishes to purchase; the android-based system will maintain the customer’s product list after scanning the price tags of the products. The bills of the individual customers would be sent to the billing server for the further verification process. The number of items added into the basket will be verified to ensure that customer has equal number of items as on the android system and his physical basket. The customer will checkout after adding the items and the payment will be made at the billing counter.

 

 

 

 

NEUROCOIN.IO

 

SALMAN ZAFAR 15k-1234 (k152880@nu.edu.pk)
ARSALAN ASHRAF 15k-1058 (k152346@nu.edu.pk)
WALEED KHAN 15k-1016 (k152380@nu.edu.pk)
MUHAMMAD ATHAR CHANGAZI 15k-3623 (k153623@nu.edu.pk)

 

SUPERVISOR

UBAID AFTAB CHAWLA (ubaid.aftab@nu.edu.pk)

 

Neurocoin.io is a platform for the buying and selling of ML models which will eliminate the distance of industry from the ML engineers who do not get enough exposure. Neurocoin.io also ensures the security of the developed model by using smart contracts on block chain and also validates the accuracy of the model developed by the ML engineer and the requirements kept by the client in the initial deal.

 

 

AAB GAARI

 

MUHAMMAD NOOR 14k-2053 (k142053@nu.edu.pk)
HAMMAD HUSSAIN 14k-2325 (k142325@nu.edu.pk)
ARFA AHMED MALLICK 14k-2060 (k142060@nu.edu.pk)
DEEDAG BALOCH 13k-2178 (k132178@nu.edu.pk)

 

SUPERVISOR

Ms. SHAHARBANO ASIF(shahar.bano@nu.edu.pk)

Mr. SHAHBAZ SIDDIQUI(shahbaz.siddiqui@nu.edu.pk)

 

Due to lack of water supply, water scarcity in different areas of Karachi is an emerging problem now a days. This project is designed to supply water for household usage in different areas of Karachi by which a person can get water without hesitation of going and search from some hydrant pumps. User will get benefit by this application as it will save the time and solve the issue of water supply.

Our project is for android users. In this application, we have three modules: the first one is for user who will place their order for water; second one is for vendors who will receive the order; and third one is for drivers who will deliver the order. Our project is based on firebase database and we are also using Google API to access the location. In future, we can add more features according to the needs of users. This application addresses the basic level need of Karachiites. As we felt that the issue dealt in this app is the most important one which have been less dealt up till now, so our app gives a full-fledge platform to tackle the issue in user friendly manner.

 

 

 

DEEP INFORMATION RETRIEVAL SYSTEM USING CONTEXT AWARE VECTOR REPRESENTATION

 

QAZI SAAD   k152851 k152851@nu.edu.pk
UMER TARIQ   k152888 k152888@nu.edu.pk
AHMED ALI   k152204 k152204@nu.edu.pk
ANAS ABID   k152413 k152413@nu.edu.pk

 

SUPERVISOR

Mr. BEHRAJ KHAN (Behraj.khan@nu.edu.pk)

 

Current retrieval systems provide good solutions towards phrase-level retrieval for fulfilling simple fact and entity centric needs. But, they struggle in concept/context based searches. The core reason this is the use of Document-term matrix or inverted index for storing documents, which can be retrieved by vector space or other language models. This means that they cannot retrieve documents from disk based on concepts, as they are indexed on terms. So, they add a layer on batch of documents returned by the key-word model to re-rank the results on conceptual level. This imposes a limitation on the architecture to retrieve only key-word based results from physical storage so even if it applies re-ranking on the fetched results a huge amount of documents will be left out which would only be relevant in terms of context. This approach also adds a layer for ranking hence would require extra amount of time, finally even conceptual ranking uses knowledge graph to conceptually understand documents that are retrieved from the physical storage. In short, it is using a translator to understand documents, so in essence it does not truly understand the content which imposes a limitation to the efficiency that can be achieved. This struggles of traditional IR reaches its maximum when tested against multilingual searches.

We propose an approach which exploits current research both in the field of Natural Language Processing (NLP) and Deep Learning (DL). Our approach will use embedding for everything from documents to query. We will then use proximity search to find nearest document in our search space. We can store documents from any language or even data in any modality in a single continuous vector space for instance image, video etc. This allows to retrieve results irrespective of language and modality.

 

PERSONALITY IDENTIFICATION THROUGH TRANSFER LEARNING USING ARTIFICIAL NEURAL NETWORK (ANN)

 

YASHARAH MIRZA 15k-3611 (k153611@nu.edu.pk)
ZOYA REHMAN 15k-2333 (k152333@nu.edu.pk)
VAJIHARAZA 15k-2135 (k152135@nu.edu.pk)

 

SUPERVISOR

Mr. BEHRAJ KHAN (behraj.khan@nu.edu.pk)

Mr. FARRUKH HASSAN SYED (farrukh.hassan@nu.edu.pk)

 

Personality is a major factor in knowing who the person really is. The better we understand people, the more effectively we can deal with people and situations. Previously, personality identification was done using dictionary based approaches and corpus based approaches.

In FYP2, we have used “transfer learning” for personality identification using convolutional neural networks (CNN) and recurrent neural network (RNN) on different datasets. We have also developed a desktop application which aims to figure out personality of users given a user dataset. Transfer learning is used to tune the model which we developed in FYP1.

 

 

QURAN COMPANION (INNOVATIVE WAY TO MEMORIZE AND REVISE QURAN)

 

ABDUL WAHAB USMANI  15k-2159 (k152159@nu.edu.pk)
ZULQARNAIN QASIM   15k-2160 (k152160@nu.edu.pk)
SHUJA ALI   15k-2424 (k152424@nu.edu.pk)
S.OSAMA HASAN   15k-3632 (k153632@nu.edu.pk)

 

SUPERVISOR

Mr. BEHRAJ KHAN  (behraj.khan@nu.edu.pk)

Dr. MUHAMMAD RAFI  (muhammad.rafi@nu.edu.pk)

 

The Huffaz (people who memorize complete Quran), after becoming Hafiz-e-Quran, enter in other fields of the world and it becomes difficult for them to maintain their Hifz. Usually, they need a partner who himself is the Hafiz to listen to their recitation of Quran and correct their mistakes. Now a days, technology is playing a great role in solving almost every kind of problem. Thus, technology may solve this problem as well. A mobile app can solve the problem. A mobile app that has complete recordings of Quran and can listen to  Hafiz’s recitation and match it with previous recordings to determines the mistakes and corrects the person’s recitation. This app surely requires help from the fields like Mobile Application, Human-Computer Interaction (HCI), Machine-Learning (ML) etc. A very important concept of “Supervised Learning” is used along with “Speech Recognition”. Basically, the app will help people (Hafiz or Non-Hafiz) in memorizing and revising the Quran. The research for this project includes collecting a number of recordings of Quran of different people (different voices) and concepts of ML to search for a way to achieve such an app and ways of implementing it.

 

 

 

 

ONLINE SKILLS, SERVICES AND BUSINESSES MARKETPLACE

 

MUHAMMAD ASAD KHAN 15k-2418 k152418@nu.edu.pk
SYED MUHAMMAD JAMILUDDIN 15k-2342 k152342@nu.edu.pk
SYED ZAIN UL EBAD 15k-2456 k152456@nu.edu.pk

 

SUPERVISOR

Ms. JAVERIA FAROOQ (javeria.farooq@nu.edu.pk)

 

The reason for the high failure rate of the freelancing projects is outsourcing through the intermediary of freelancer’s marketplace, which causes wasted resources. Projects done do not always satisfy the quality, time and budget constraints of the client. Online skills, services and business (OSSB) aims to provide a quality in projects, and trustworthy good relationship between freelancers and clients by providing the new feature of the review based feedback. In order to provide accurate review, all the reviews will be rated using rule based sentiment analysis, which will provide transparency in exploring the trustworthy freelancers. Additionally, OOSB provides bidding opportunity to freelancers keeping in mind the budget constraints of the client. Our aim is to encourage freelancing market, which will benefit the small and medium enterprise.

 

 

 

 

PRICECOM

 

M. WAQAR ALI ANSARI 15k-2482 (K152482@nu.edu.pk)
HAMZA ABDUL MATEEN 15k-3620 (K153620@nu.edu.pk)
SALAR QURESHI 15k-2468 (K152468@nu.edu.pk)
NIRMAL HASSAN RIZVI 15k-3637 (K153637@nu.edu.pk)

 

SUPERVISOR

Mr. AHSAN SHAH (ahsan.shah@nu.edu.pk)

 

Searching through multiple e-commerce sites is a dilemma. A single price comparison platform with product based and user based recommendation was much needed. PRICECOM, a solution to modern e-commerce product searching problem, gives users a single platform to search from, for products they are willing to buy or get information for. PRICECOM aims to solve few present e-commerce problems faced by users in their day-to-day visit to different e-commerce platforms. PRICECOM is based on an Android application and a website on the front-end and is based on python for scrapping, node Js for recommendation system and MySQL for database as backend.

 

 

 

 

SMART BIDDING APPLICATION (BidX)

 

MUHAMMAD HASHIR15k-2345 (k152345@nu.edu.pk)
MOHAMMAD WAJHI UDDIN 15k-2823 (k152823@nu.edu.pk)
TALHA SHAHAB 15k-2153 (k152153@nu.edu.pk)
MUHAMMAD MOIN RAZA 15k-2319 (k152319@nu.edu.pk)

 

SUPERVISOR

Mr. AHSAN SHAH (ahsan.shah@nu.edu.pk)

 

In the modern era of online exchange systems, the value of time is an important factor. Due to the advancements in technology and limitations of time, the basic model for buying and selling products that was bidding was left behind. The seller was not able to purchase a product on time with the real worth of the product and the buyer was getting the product in either expensive price or with delay. In order to devise a solution, our team came up with the idea of Online Smart Bidding Application (BidX) that will provide users a native platform (Android/IOS) to buy and sell products within a specific time constraint and efficiently. We have used Restful API’s for the purpose of machine learning algorithms to make this app even smarter than the previously build applications. We are providing users the predicted price of their product based on the ML algorithms so that they will not be in loss, and also using ML to filter fake ads from the reach of our user.

 

 

 

DRIVE LEARNING APP

LEARN-PRACTICE-DRIVE

 

MAWIZ AKRAMA  15k-2248 (k152248@nu.edu.pk)
HAMNA MALLICK  15k-2208 (k152208@nu.edu.pk)
HASSAN ALI   15k-2253 (k152253@nu.edu.pk)
HAMZA SHAMS   15k-2830 (k152830@nu.edu.pk)

 

SUPERVISOR

Mr. SHOAIB RAZA (shoaib.raza@nu.edu.pk)

Mr. SAYED YOUSUF (sayedyousaf@yahoo.com)

 

Learning to drive a car is an important skill that everyone should and wants to learn. The problem that arises when someone decides to learn to drive is where/who should they learn it from. The go-to option for most people is driving schools. We over viewed some driving schools in Karachi and discovered that these institutes charge around 10k to 15k for a driving course of 4-5 weeks, which includes some theory classes and 1 hour driving lessons on alternative days. Paying a large-sum of money at the beginning of the lesson can be a burden for some people, so the proposed solution for this is an app that connects customers willing to learn to drive with driving institutes in their area and let them choose the institute that they prefer. Inexperienced drivers would be connected to licensed driving instructors from the institute that they choose at any time that suits the learner. The learners would be able to pay with cash for each lesson separately instead of paying a lump sum fees. The app will help amateur drivers develop their theoretical knowledge about driving by providing structured lessons and step by step guides covering all the information from set off to advance every day driving. For practical driving lessons learner drivers will browse through the list of driving institutes that operate in their area, select an institute and then request for an instructor by booking a lesson. Learners will be able to filter their request by car type preferences, gender preference for the instructor and fare estimation (calculation showing “starting from base fare”). Instructors will then pick the learners for lessons and the entire lessons will be tracked on the app. The learners will be charged for each session on the basis of distance travelled, time and base fare (the amount of tuition cost fixed by the institute) rate i.e. Distance charge + time charge + tuition cost fixed by institute (for elaboration only). After each lesson is complete the instructors will log the progress of the lesson in the app so the next time a learner takes a lesson from a new instructor, they know where to start the next lesson from.

 

 

MEALS ON WHEELS

 

SYED JAFFAR KAZMI    15k-3629 (K153629@nu.edu.pk)
KARAN KUMAR    15k-2111 (K152111@nu.edu.pk)
ALI ASADULLAH    14k-2157 (K142157@nu.edu.pk)

 

SUPERVISOR

Mr. SHOAIB RAZA (shoaib.raza@nu.edu.pk)

Mr. SAYED YOSOUF(sayed.yosuf@nu.edu.pk)

 

MEALS ON WHEEL is the food delivery application with the potential to reach the target audience consists of students and professionals in an effective manner and ensures the delivery of fresh and hygienic foods to them at a tapping distance. Just like every other food delivering apps Meals on Wheels delivers food to customers, but it has some uniqueness: first the delivered food is homemade- hygienic and low on budget; secondly, the application tracks the rider’s location and time, so that customers are fully aware of delivery time. The application designed to cater the demands of the students who travels miles away from their home in search of better education among many other hurdles they face the one which is of great significance is finding the best food in town in terms of taste, readiness and quality. Other than students there are office workers who also prefer to order homemade foods at their workplaces. Another great aspect of this application is that it creates opportunities for people to earn in a right way especially to housewives, widows and needy women, who can start their businesses by selling food from home without facing the hassles of marketing, endorsing and other business matters. The application will consist of three interfaces: Customer, Chef and Rider. Customers can search and order food from different Chefs with the option of selecting the nearest Chef available. Chefs can update their menu and food availability, and Riders gets delivery notifications and details.

 

 

 

 

PAKISTAN SIGN LANGUAGE (PSL) CHAT PROJECT

 

MAAZ BIN AAMIR 15k-3641 (K153641@nu.edu.pk)
MURTAZA ZAIDI 15k-3604 (K153604@nu.edu.pk)
TALHA MEHMOOD 15k-1016 (K152349@nu.edu.pk)

 

SUPERVISOR

Mr. SHOAIB RAZA   (shoaib.raza@nu.edu.pk)

ENGR.ABDUL RAHMAN (abdulrahman@nu.edu.pk)

 

The human disabilities differs from person to person, and there are three categories of physical disability (Blind, Deaf and dumb, and crippled). Our proposed solution is focused on to develop a communication approach for the conversion of text and signs as visual representation. The design is focused on the objective of reducing the communication gap between normal people and impaired people. The proposed design has two applications, one used for disabled person, and other used to normal human. Two applications are designed, one for deaf-dumb chat box who select the sign(s) and retrieve from database the meaning in alphabetic form, and send it to other in text format. The other application its used by normal human whom write the message and send it to first application on other PC via network and there will be text to sign conversion which will show signs to first application.

.

Keywords: Pakistan sign language, server and client.

 

 

 

AI DOC

 

                                     SAMEERUDDIN      15k-2826                        (k152826@nu.edu.pk)
                                        SHAHAN ALI       15k-2239                        (k152239@nu.edu.pk)
                        MUHAMMAD FAIZAN      15k-2105                        (k152105@nu.edu.pk)

 

 

 

SUPERVISOR

Mr. ZESHAN KHAN  (zeshan.khan@nu.edu.pk)

Mr. FARRUKH HASAN SYED  (farrukh.hassan@nu.edu.pk)

 

A nation is considered to be strong if they have major experts in the field of medical science. As we know that computer science is making progress day by day, so now it is our responsibility to take our medical field into latest technology era by applying artificial intelligence. Our proposed system will diagnose disease with the help of machine learning tools.  The goal is to assist doctor in every possible phase of diagnosing disease through expert system. These systems work on enriched data in which we can apply rule based algorithms.

 

 

 

 

 

 

 

ECG INTERPRETATION AND RHYTHM ANALYSIS

 

M. HAMZA ISMAIL 15k-2843 (k152843@nu.edu.pk)
HUZAIFA EJAZ  15k-2330 (k152330@nu.edu.pk)
HAMZA AHMED 15k-2302 (k152302@nu.edu.pk)

 

SUPERVISOR

Mr. ZESHAN KHAN (zeshan.khan@nu.edu.pk)

 

The Electrocardiogram (ECG) is an established technique in cardiology for the analysis of cardiac condition of the patients. An ECG (electrocardiogram) is an important diagnostic tool for the assessment of cardiac arrhythmias in clinical routine. The clinical doctors analyze the ECG graph and on the basis of the calculations and observations, he diagnoses some heart diseases and because of that the accuracy to analyze the diseases is not perfect. In current medical routine, careful study of the ECG by expert cardiologists is necessary for the diagnosis of life threatening cardiac arrhythmias. However, automatic classification of cardiac arrhythmias can both provide objective diagnostic results and save time for the cardiologists. The technical aspect of this project is to extract the values PQRST by using the image processing algorithms of OpenCV in python so from the technical aspect this product based project is feasible but there is no well-known tool available to analyze the ECG graph and diagnose diseases. The aim of this project is to develop a diagnostic system which assists cardiologists as well as clinical doctors by providing intelligent, cost effective and time saving by diagnosing the values of P, Q, R, S, and T in order to identify ECG arrhythmia diagnostics. To achieve this goal, image processing and pattern recognition methods are implemented on the ECG images of MIT-BIH dataset in order to test the system. The Basic idea is to automate the process of ECG analysis and diagnosis of heart diseases with more accurate and precise.

 

 

 

 

 

 

 

 

 

Department Of Electrical Engineering

 

UNDER GROUND CABLE FAULT LOCATOR

 

SYED HAMZA ARSHAD 15K-2383 (k152383@nu.edu.pk)
HASSAN UL HAQ BABAR 15K-2396 (k152396@nu.edu.pk)
MUHAMMAD UZAIR YOUSUF 15K-2452 (k152452@nu.edu.pk)
ALI ASHAR 15K-2406 (k152406@nu.edu.pk)

 

SUPERVISOR

ENGR. AQIB NOOR (aqib.noor@nu.edu.pk)

 

 

Robotics is one of the fastest growing engineering fields of today. Robots are designed to remove the human factor from intensive labor and dangerous work and also to act in an inaccessible environment where human interface is not possible. Till last decade, a million miles of cables were threaded across the country. Now, gradually the process of laying the cables underground has started, which is much better than the earlier method. Underground cables are not affected by any adverse weather condition like pollution, heavy rainfall, snow and storm, etc. However, when any problem occurs in cable, it is very difficult to find the exact location of the fault that has occurred. Today, the world is progressing towards innovation and latest technology, so keeping this in mind, a project has been designed that will detect the exact location of the fault. In this project a special purpose robot has been designed that will move above the underground cable and detect the fault. This task has been performed by the use of different sensors to detect the fault. In addition, camera, GSM, GPS, IR and, proximity sensors are used which helps in finding the cable fault. When a cable fault occurs, it causes big problems and a speedy solution is very important. To overcome this problem, a prototype has been designed that will identify the exact location of the fault.

 

 

SMART IRRIGATION SYSTEM

 

ZUBAIR AHMED 15K-2382 (k152382@nu.edu.pk)
ABDUL RAUF 15K-2381 (k152381@nu.edu.pk)
NAJEEBULLAH 15K-2471 (k152471@nu.edu.pk)
MOHAMMAD HAMZA 15K-2357 (k152357@nu.edu.pk)

 

SUPERVISOR

ENGR. SYED KASHAN HUSSAIN (kashan.hussain@nu.edu.pk)

ENGR. UMAIR NAEEM (umair.naeem@nu.edu.pk)

 

 

Technology is continuously and rapidly changing. It is used in every field either in the field of computer data handling, defense or any other field. In current situations, in the field of irrigation there should be use of technology for better growth and management of water in appropriate way. In the developing world the water management is improving in its efficiency.

In the irrigation, the water is necessary for land or soil. Due to water scarcity, irrigation system needs to be improved. The automatic irrigation system helps in the proper growing of the crops by the water management as it maintains the water level which helps in improving the irrigation. An automated smart irrigation system has been proposed for efficient water management. In modern world, the Smart irrigation system helps to reduce the number of workers and cost in the field, increases the crop growth and controls the wastage of water and time. In the field of irrigation use of proper methods are important for economic growth. The automated irrigation system maintains the field automatically. It checks the water requirement of the soil in each period. The automated irrigation system is mainly consisted of simple sensors. The product is best for those areas where there is water deficit. The soil moisture is difficult to measure and to maintain its efficiency. For that reason, the smart automated irrigation system is best to solve this problem.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AUTOMATED AIR FIELD CONTROLLING AND MONITORING SYSTEM

 

MUHAMMAD WAHAJ BIN SHABAZ 15K-2394 (k152394@nu.edu.pk)
SYED SAAD ALI AFFANDI 15K-2410 (k152410@nu.edu.pk)
MUSTAFA AHMED 15K-2455 (k152455@nu.edu.pk)
AYESHA MUKHTAR 15K-2508 (k152508@nu.edu.pk)

 

SUPERVISOR

ENGR. MUHAMMAD AHSAN KHAN (muhammad.ahsan@nu.edu.pk)

ENGR. MAHAM GHAURI (maham.ghauri@nu.edu.pk)

 

 

Worldwide trends in the aviation industry indicate that monitoring and controlling airfield lighting systems increases operational efficiency and improves the safety in air-traffic control and aircrafts. Airfield ground lighting systems are responsible of highlighting the runway, giving visual reference of speed and alignment to aircrafts in the final stages of approaching, landing, and taxiing operations that could easily be overlooked by air travelers, hence play a vital role in safe air travel. We considered the history of airfield lighting, and took a step towards a potential future development. The motive of this project is to control the airfield runway lighting system automatically as well as manually that includes factors affecting lighting which may endanger the safety of aircraft.

The automatic control of lighting system is controlled by Current Controlled Regulator-CCR. The main purpose of CCR is to control the lighting system and the intensity of the lights as per requirement of pilot and weather condition. This project is a prototype that concluded the implementation, control and supervision of airport lighting system that further aims to build a Visual Docking Guidance System for the Aircraft to safely Park at the designated point. This Project Aims to facilitate and improve the system installed at the airport keeping in mind the latest technologies and techniques which may help in improvising the system and reducing the cost or the chances of errors.

 

 

 

 

IOT BASED SMART ENERGY GRID

 

ABDUL FATAH 14K-2433 (k142433@nu.edu.pk)
ALLAH BAKHSH 14K-2511 (k142511@nu.edu.pk)
KASHIF SAEED 14K-2444 (k142444@nu.edu.pk)

 

SUPERVISOR

ENGR. MUHAMMAD JUNAID RABBANI (junaid.rabbani@nu.edu.pk)

ENGR. ASHIR MEHBOOB (aashir.mahboob@nu.edu.pk)

 

 

The electricity is supplied to the households via intermediate controlled power transmission hubs known as Electricity Grid .The problem arises due to the failure of the electricity grid leading to the black out of an entire area which receives supply from that particular grid. Therefore, this project aims to resolve these issues using IOT as the mean of communication, and also to tackle various other issues which the smart system can deal with to avoid unnecessary losses to the Energy producers. The Grid is based on controller that distinct activities of the system. The system communicates over internet by using Wi-Fi technology. Two bulbs are used; one to show as a valid consumer, and other to indicate an invalid consumer. The foremost thing is to facilitate is the re-connection of transmission line to active grid. If any Grid becomes faulty and there is another Grid, the system switches the Transmission Lines towards that Grid; thus, facilitating uninterrupted electricity supply to that particular region whose Grid went OFF. This information is updated on web site where the authorities can login and can view the updates. Apart from monitoring, the project has capabilities to monitor energy consumption and also detect theft of electricity. The amount of electricity consumed and the estimated cost get updated on web site along with the Energy Grid information. There are two switches that indicates energy theft. Energy theft are of two types; internal and external. The internal theft is detected when the energy consummation is more on the output of the energy meter than input. If the case is reversed, it is external theft. Both are shown by two loads. Theft conditions are simulated in the system. Each time, it simulates a theft condition, it will notify the authorities over the IOT interface. In this way, the Smart Energy Grid project makes sure that the electricity supply is continuous and helps in maintaining an updated record of consumption and theft information which is quite a valuable information for the energy.

 

 

 

 

SOLAR HYBRID CAR

 

WASIQ SULEMAN 15K-2500 (k152500@nu.edu.pk)
JAHANZEB ASAD 15K-2372 (k152372@nu.edu.pk)
MOMINA KAMAL 15K-2470 (k152470@nu.edu.pk)
MAIRA FAROOQUI 15K-2504 (k152504@nu.edu.pk)

 

SUPERVISOR

ENGR. AQIB NOOR (aqib.noor@nu.edu.pk)

ENGR. BILAL YOUSUF (bilal.yousuf@nu.edu.pk)

 

 

Combustion engine vehicles are responsible for different kinds of pollutions as they are the major contributors of carbon dioxide production, fossil fuel consumption, thus resulting in major climate changes. Solar power for transportation can control pollution and also minimize the impact of Global warming. This project is articulated to propose a Solar Electric Powered Hybrid Vehicle (SEPHV) system which solves the major problems of fuel and supports Green energy. An electric solar car functions on the principle of photovoltaic cells that transform sunlight into energy; Lithium-ion batteries are used in Hybrid cars for reserving the Electric Power generated from Photovoltaic Cells. This Solar Electrical vehicle can charge itself from both solar and electric power and is more fuel-efficient with regard to the Green-house gases emission control.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IOT BASED SMART BANK SECURITY AND SURVEILLANCE SYSTEM

 

SYED ZAIN UL ABIDEEN 15K-2435 (k152435@nu.edu.pk)
SYED FARAZ ALI 15K-2439 (k152439@nu.edu.pk)
SYED TAHA ZAIDI 15K-2376 (k152376@nu.edu.pk)
HAMZA SHAIKH 15K-2491 (k152491@nu.edu.pk)

 

SUPERVISOR

ENGR QURATULAIN SOHAIL (quratulain.sohail@nu.edu.pk)

 

 

Due to increase in the cases of theft and risk of intrusion, security has become one of the integral issues to ensure the safety of the resources. Security for banks and lockers is necessary and to safeguard the bank, a security system using different sensors (smoke detector, PIR sensor) is developed to detect the movement and abnormal activities taking place inside the bank. Fingerprint sensor is also installed which protects the bank from unauthorized entry. Furthermore, Smart camera system is being develop for the surveillance of the bank and the whole system is integrated by Raspberry Pi which sent live streaming of bank through IOT (Internet of Things)

 

 

 

LASER ENGRAVING MACHINE USING DIFFERENT CONTROLLERS

 

SHAKEEL AHMED 15K-2447 (k152447@nu.edu.pk)
TAHIR ZAMAN 15K-2454 (k152454@nu.edu.pk)
SHUMAIL IQBAL 15K-2451 (k152451@nu.edu.pk)
HUMAYUN IQBAL 15K-2437 (k152437@nu.edu.pk)

 

SUPERVISOR

ENGINEER SABOOR KHAN (abdul.saboor@nu.edu.pk)

 

 

In laser engraving, the laser beam burns the top layer of the surface to be engraved. The burnt area is left uncolored which makes it appear different from the surrounding surface. In laser engraving, the laser beam has to engrave the surface that can be achieved by prolonging the beam on a particular area for a long period of time; duration is decided on the strength of material to be engraved. We are still collecting data from different resources to know this topic more in depth. In this project, we have decided to develop working model of laser engraving, and engraving machine. This machine is very useful in our department since it is a kind of rapid prototyping machine. It can engrave sheets and materials to produce desired shape and patterns. Till now we have collected information on laser engraving, laser, and what engraving is. Laser engraving is different from laser engraving, in case of engraving very low intensity laser torch is used as compared to laser engraving. We have provided a brief introduction on laser, how it works, what are its properties, how to generate it and how to control its intensity. The machine is made using a 3000mW red laser. It may engrave the wood and other material surely. In order to simulate each part of the machine, simulation analyses are performed in CAD software ‘CATIA V6’. It was helpful for remodeling the moving bed or the job holder if any errors found during the simulation. Additionally, experiments are performed to develop laser engraving machine.

 

 

 

 

COMPUTER VISION BASED COMMUNICATION ASSISTANCE FOR AUDIBLY IMPAIRED PEOPLE

 

MUHAMMAD UMAR JAHANGIR 15K-2389 (k152389@nu.edu.pk)
ASIM MEHDI 15K-2401 (k152401@nu.edu.pk)
TOOBA ATHER 15K-2399 (k152399@nu.edu.pk)
MUHAMMAD UMER FAROOQ 15K-2364 (k152364@nu.edu.pk)

 

SUPERVISOR

Dr. HAIDER MEHDI (haider.mehdi@nu.edu.pk)

 

 

To communicate with a hearing impaired individual, a computer-vision based project is developed that will help to communicate via gestures by vocalizing sign language. This is done via a raspberry pi integrated with a camera, observing for a gesture and giving a vocal output to help a deaf person to achieve clear communication with someone unfamiliar with sign language. Our device (Raspberry Pi with Camera) will recognize a specific hand gesture and will give audio output for that gesture. We have used pre trained models for American Sign Language. Alphabets individually are not that robust, so we went for scenario-based approach, in this case a milk shop. We have used resnet50 pre trained models on PyTorch using Deep Learning. Our hardware looks like normal glasses on which we have fixed Raspberry Pi, Camera and Speaker with Power Bank. We are using glasses because camera will easily detect gestures from that view.

 

 

 

THE SMART HELMET

 

ABDUL AHAD 15K-2453 (k152453@nu.edu.pk)
MOHAMMAD FAHAD 15K-2426 (k152426@nu.edu.pk)
UMER AHMED 15K-2436 (k152436@nu.edu.pk)

 

SUPERVISOR

ENGR. QURATULAIN SOHAIL (quratulain.sohail@nu.edu.pk)

ENGR. MAHAM GHAURI (maham.ghauri@nu.edu.pk)

 

 

A smart helmet is an uncommon thought which makes motorcycle driving more safe and secure than before. The idea of this project was to create a safer environment for the two-wheelers as most accidents that occur cause severe damage to them. The Smart Helmet will become a necessity for the riders to wear as without wearing them, they won’t be able to ignite the bike.

The implementation uses the technology of GSM (Global System for Mobile) and GPS (Global Positioning System). Sensor is placed on the vehicle which is connected to microcontroller board. So, whenever the rider is involved in an accident, the sensor senses and signals to the microcontroller. It then draws out the exact GPS data using the module which is interfaced to it and then sends a message to the family members with the GPS coordinates and the map link. The brain of the whole system is the Arduino UNO.

 

 

IOT BASED SMART MIRROR

 

MUHAMMAD ANAS 15K-2403 (k152403@nu.edu.pk)
SAAD AYUBI 15K-2469 (k152469@nu.edu.pk)
SAAD RASHID 15K-2499 (k152499@nu.edu.pk)
ABDUL MANAN 15K-2518 (k152518@nu.edu.pk)

 

SUPERVISOR

ENGR. KASHAN HUSSAIN (kashan.hussain@nu.edu.pk)

ENGR. AAMIR ALI DAYO (aamir.ali@nu.edu.pk)

 

 

In this project, we are focusing on making a use of simple mirror in a more avant-garde way. For making the mirror smart, we have used the concept of Internet of Things (IoT). Internet of things gives ability to an object to transfer data over a network without the interference of human or we can say that the IoT transfer our lives by connecting us more easily to information and other people in the virtual world. Mobile phones have become smart phones, and there is no end of things becoming smarter day by day so now it is time to make mirror smart. For making the mirror smart, we have used the voice assistance which will act as interference between mirror and user. When it turns OFF, it will reflect the image of the person. But when it is turned ON, more informational things can be seen like current weather, time, date, news etc. It will not only give reviews on the looks, but it will also be awake on the name you will give to your mirror like hello ALEX or hello pi etc. we have used Raspberry pi 3 as a controller for making this possible. Materials used in smart mirror are easily available and affordable, and all the soft-wear used in this project are open source and required small amount of processing programs that will consume limited amount of internet while running. All the components are integrated in a wooden frame which makes the project compact, and plug and play device. Smart mirror is the proof of advancement of technology in the field of artificial intelligence surely smart mirror will modify our lives and way of living.

 

 

 

SIX AXIS ARTICULATED ROBOTIC ARM FOR INDUSTRIAL APPLICATION

 

 

MUSADIQ ABBAS 15K-2479 (k152479@nu.edu.pk)
ARSALAN ABBAS 15K-2405 (k152405@nu.edu.pk)
REHAN ALTAF 15K-2501 (k152501@nu.edu.pk)
ZUBIA KHAN 15K-2402 (k152402@nu.edu.pk)

 

Supervisor

ENGR. MUHAMMAD AHSAN KHAN (Muhammad.ahsan@nu.edu.pk)

 

 

Human body can be considered a very complex machine which is a perfect combination of joints that have the ability to reach maximum degree of movement. Technology is advancing on an unprecedented rate and robots have even replaced humans in countless fields including linguistics, engineering, medicine etc. This revolution is in part to the incredible accuracy and countless man hours that robots can engage for, especially when compared to humans. The shortcoming of machines is that they are restricted to specific tasks and lack the dynamic skills that a human does. Our objective is to create one such robust machine, capable of multiple tasks, resembling the work and adaptability of a human. Furthermore, we aim to exceed the limitations of a human hand and produce results that would be otherwise unattainable by a human. However, as our basis, our project derives its improvements by observing the mechanics and working of the human body.

 

 

IOT-BASED GROCERY MANAGEMENT SYSTEM

 

MUHAMMAD ASAD KHAN 15K-2434 (k152434@nu.edu.pk)
HASSAN MANSOOR 15K-2443 (k152443@nu.edu.pk)
MUHAMMAD HAYYAN BIN SHAHID 15K-2369 (k152369@nu.edu.pk)
UZAIR SHAFIQUE 15K-2422 (k152422@nu.edu.pk)

 

SUPERVISOR

Dr. BURHAN KHAN (burhan.khan@nu.edu.pk)

ENGR. AAMIR ALI DAYO (aamir.ali@nu.edu.pk)

 

 

The Internet of Things (IoT) is about connecting things to the internet and enabling them to share their data. It is a network of physical objects which collect and transmit information related to those objects. This network contains embedded sensors, actuators, software and connectivity to the internet. The IoT has impressively digitized the physical world. It will help to boost the economy as much as $11 trillion a year in 2025. Basically, IoT is about to make the system intelligent and connect them with internet so that users can control and monitor the things easily on their electronic devices. The world is getting digitalized and every work now a days can be done remotely. To contribute in the smart world, we want to provide people with the IoT-based Grocery Management System that consists of a Smart Refrigerator and Smart Pantry. Smart Pantry can be used to show data about vegetables, pulses, spices, rice, flour etc. Smart Refrigerator will show all the contents placed inside it. User will be able to access all this data via a dedicated web application from his/her mobile, tablet, and computer from anywhere in the world. With the permission of user, this IoT-based Grocery Management System, will place an automatic online order to replenish the finished food items. User will be notified via email about the placement of this order.

 

 

 

ELECTRONIC VOTING SYSTEM

 

M. SALMAN MASOOD 15K-2365 (k152365@nu.edu.pk)
HAMZA BIN EHTESHAM 15K-2432 (k152432@nu.edu.pk)
M. FAHAD MANSOOR 15K-2467 (k152467@nu.edu.pk)
AFIFA RAZA 15K-2488 (k152488@nu.edu.pk)

 

SUPERVISOR

ENGR. HARIS MOHSIN (haris.mohsin@nu.edu.pk)

ENGR. UMAIR NAEEM (umair.naeem@nu.edu.pk)

 

 

Pakistan is a democratic country and voting is an important part of the democratic process of any country. The result of the polling is concluded within one day and the supreme power goes to the winning party. But in Pakistan, it is not as simple as we think. Polling does not go properly and results into ballot rigging. Moreover, unwanted use of manpower is widely observed during the general elections. For such type of flaws, we have come up with a solution of introducing a unique electronic voting system that will be efficient, automated, user friendly, and portable. Our system includes the e-voting machine, auto-counter ballot box which will be used for the dual verification of total votes, CCTV monitoring and saving the poll data to be displayed just after the end of polling time. The whole process of our system will work in such a way that a voter scans his/her thumb print on the finger print sensor, casts his/her own personal vote, drops his/her generated slip into the ballot box and receives a confirmation message on his/her mobile phone of successfully casting vote. Our whole system will be controlled by Raspberry Pi 3 B+. However, the Auto-Counter Ballot Box will be controlled separately by an Arduino UNO.

 

 

 

FORMATION CONTROL OF MULTIPLE UAVS

 

JAMSHAID ZAIDI 15K-2427 (k152427@nu.edu.pk)
HAMZA FAROOQ 15K-2494 (k152494@nu.edu.pk)
ZOBIA NADEEM 15K-2449 (k152449@nu.edu.pk)
HASHER WASEEM 15K-2484 (k152484@nu.edu.pk)

 

SUPERVISOR

ENGR. HARIS MOHSIN (haris.mohsin@nu.edu.pk)

ENGR. MUHAMMAD BILAL YOUSUF (bilal.yousuf@nu.edu.pk)

 

 

Formation control techniques are burgeoning vis-a-vis other operations involving single UAV. This can be credited to the wide array of applications involved with the use of multiple UAVs for a task rather than a single. The swarming technique for the coordinated control of UAVs. The Matrix decomposition method is used in this regards and appropriate gain values are assigned to the eigenvector in order to keep the UAVs in a trouble-free and well-ordered formation. The communication topology applied is straightforward where both drones are assigned with different frequencies in order to avoid any communication losses that the UAV may encounter if assigned with the same frequency value.

The hardware prototype involves two drones one of which is designated as a master while the other one is designated as the slave. The UAV is programmed in python and the controller used in Ardu Pilot Mega 2.8. Four brush-less carbon motors are used that are mounted on the end of the X-frame. Two telemetries of different frequencies are used. The frequency of 415 MHz is assigned to the master drone whereas the frequency of 915 MHz is assigned to slave drone.  The telemetry basically has two modules i.e. Air module and Ground Module. The air modules of both telemetries are mounted onto their respective UAVs. The master drone sends the command to the ground station setup that is linked to the IDE Mission Planner. The ground stations of both drones are synchronized and hence the command is sent to the air module of slave drone via ground station.

 

SMART ATMOSPHERIC WATER GENERATION UNIT

 

DANISH NOOR 15K-2490 (k152490@nu.edu.pk)
SHAMS MURTAZA 15K-2417 (k152417@nu.edu.pk)
HAMZA MUFTI 15K-2377 (k152377@nu.edu.pk)
SHARJEEL DURRANI 15K-2359 (k152359@nu.edu.pk)

 

SUPERVISOR

ENGR. SYED ASIM MAHMOOD (asim.mahmood@nu.edu.pk)

ENGR. UMAIR NAEEM (umair.naeem@nu.edu.pk)

 

 

About a billion of people around the globe are affected due to scarcity of water. According to the research it is found that the areas that are water scarce are mostly the regions near the coast even having the sea nearby, there is shortage of drinkable water. The desalinization is a solution to this problem in such areas. This is not a cheap method to extract water and fulfill the needs. The proposed method is the cheap solution to this water related problem. The atmospheric water generator requires the humid air to extract water out of it that is sufficiently available in the coastal regions making it a good source of obtaining water.

 

 

SMART JACKET FOR SAFETY OF COAL MINERS

 

SYED GHULAM ABBAS ZAIDI 14K-2397 (k142397@nu.edu.pk )
M. WALEED GHAURI 14K-2522 (k142522@nu.edu.pk )
KHUDAIJA AFSAR 14K-2527    (k142527@nu.edu.pk )
WALEED UDDIN RAZI 14K-2418    (k142418@nu.edu.pk )

 

SUPERVISOR

Dr. SYED ATIF SALEEM (syed.saleem@nu.edu.pk)

ENGR. UMAIR NAEEM (umair.naeem@nu.edu.pk)

 

 

“Smart Jacket for the Safety of Coal Miners” proposes a cheap and very efficient assistive system for the miners who lose their lives in the coal mines. In this project, basic operation to detect some fundamental information including miners vital signs, depth of the miner, pressure in the mine, temperature under the mine, gas level and its location. In this device we have proposed an engineering cap reshaped and re designed by fixing some of the sensors that will be placed to detect the parameters we require. This signal will be later on transmitted by wirelessly using ESP8266 Wi-Fi module and GSM module to the IP of the router. The purpose of generating this method statement is to define the procedure step by step to implement the cheap but efficient device for the coal miners; moreover, to ensure that the device installation work execution compiles with the project requirement and serves the intended function to satisfactory level.

 

 

 

AUTONOMOUS UNDERWATER VEHICLE ROBOT

 

KHURRAM MUMTAZ 15K-2390 (k152390@nu.edu.pk)
NAYAB MEHER 15K-2391 (k152391@nu.edu.pk)
SANA SAEED 15K-2408 (k152408@nu.edu.pk)
OSAMA FAISAL 15K-2510 (k152510@nu.edu.pk)

 

SUPERVISOR

ENGR. SYED ASIM MEHMOOD (asim.mahmood@nu.edu.pk)

ENGR. MUHAMMAD BILAL YOUSUF (bilal.yousuf@nu.edu.pk)

 

 

In this paper, the problem of tracking an Autonomous Underwater Vehicle is discussed. The main idea is to interface the hardware with the simulated results. The controller is designed in three planes namely Yaw, Pitch and Depth. PID is used as the control scheme for trajectory tracking. This paper suggests solution particularly to the problems of controlling the heading of AUV in a horizontal and stern plane. The simulation result has been shown to prove the performance of the controller. Based on the simulation result, we designed hardware for the experimental purpose to solve the practical problems like monitoring and rescue operations.

 

 

 

 

 

 

 

 

 

 

 

Department Of Management Sciences

 

 

CONSUMER BUYING BEHAVIOR ON E COMMERCE THE CASE OF VEGETABLES

 

 

ANUM KHAN 15K-2704

 

(k152704@nu.edu.pk)

MOHAMMAD DANIYAL15K-2719 (k152719@nu.edu.pk)
ABDUL BASIT 15K-2723 (k152723@nu.edu.pk)

 

 

SUPERVISOR

MUHAMMAD AHSAN (muhammad.ahsan@nu.edu.pk)

 

 

E-commerce is growing at an unprecedented rate in Pakistan. However, in the context of vegetables, the growth rate of e-commerce appears to be dormant. This research explores the relationship among consumer buying behavior and variables including price, quality, convenience and trust issues pertaining to e-commerce. The research is quantitative in nature utilizing primary data which is collected using questionnaires and analyzed using Regression as a test to explore the relationship among the aforementioned variables.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BOARD OF DIRECTORS AND FIRM’S PERFORMANCE

 

 

BAKHTAWAR BALOCH 14K-2742

 

(k152742@nu.edu.pk)

ROZINA JAMALI 15K-2710 (k152710@nu.edu.pk)
SHAZIA SAHEB 15K-2716 (k152716@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

This research paper examines the relation between the effectiveness of board of director and the financial performance of the firms that are listed in Pakistan stock exchange (PSX).The purpose of the paper is to highlight the attribute of top management that impacts positively and negatively on the profitability of the firm, and it is also related with the economy of the country. The study used linear regression analysis over the period of five years from 2012 -2017 using the sample of 30 companies of cement and steel sector that are listed on PSX. The impact of board of director on firm performance is analyzed by seven characteristics of board: Non executive director, age of director, female presence in board, CEO duality, qualification of director, board meetings and board size and the performance of firm measured by financial indicators that are ROA, ROE and Leverage.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EFFECTS OF CAPITAL STRUCTURE ON THE PERFORMANCE OF MANUFACTURING FIRMS IN PAKISTAN

 

 

HASAN RAZA 15k-2714

 

(k152714@nu.edu.pk)

ARUBA 15k-2711 (k152711@nu.edu.pk)
ALI ZAIN 15k-2706 (k152706@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

The impact of capital structure is widely argued that whether capital structure has any effect on firm performance or not. The Modigliani and Miller theory of capital structure suggests no relation. On contrary, trade off theory argues that on some extent, financing by the Debt saves tax and benefits organization. The aim of our research is to determine the effects of capital structure on manufacturing companies of Pakistan. The Control variables are short term debts and long term debts; the dependent variable we have used is Return on Equity (ROE). The needed data for the research is collected from the archive, annual reports of companies. The sample contains 45 manufacturing firms listed on Pakistan stock exchange since 2014. The companies are randomly selected from different manufacturing sector i.e. cement, fertilizer, chemical, and automobiles. Correlation method is used to identify the relationship between the Debts and the Return on Equity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IMPACT OF RETAIL LAYOUT, DESIGN AND LOCATION ON CONSUMER BUYING BEHAVIOUR

 

 

RAJ SINGH 15k-2709

 

(k152709@nu.edu.pk)

SARANG MOGHAL 15k-2708 (k152708@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

The purpose of our research is to find the gap of location in retail supermarkets along with other factors that affect retail sale and contribute in retail sales are also considered. Major focus of our research is retail location (i.e. Traffic within store and Parking) and also to focus on retail design and retail layout. This research also identifies the relationship between dependent and independent variables which are buyer buying behavior (Demographics) and Instore traffic respectively. This is the primary research and data will be collected through questionnaire and for analysis SPSS software will be used. In this study samples of 400 respondents will be observed to measure the impact of purposed variables.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EXPLORING STUDENT CHOICE CRITERIA FOR SELECTING BUSINESS SCHOOLS IN KARACHI

 

 

SHAZRA MANZOOR 15k-2702

 

(k152702@nu.edu.pk)

ARFA QAMAR 15k-2712 (k152712@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

This paper investigates the selection criteria for choosing business school in Pakistan. In a student’s life, choice of institution which he/she is going to attend after the FSC/intermediate/A levels is of critical nature. Because it includes so many decisions before getting their admission done in preferred university, in this way many factors influence this choice made by the students. These may include the perceptions of people about universities, parental choice, and different factors such as location, distance, quality of education, faculty, fee structure, campus and facilities, recognition in job market, and alumni strength which students generally develop on their life. It is because every student wants to enroll in good and reputed university which can help them to target their future goals. For students the selection of an institution to obtain a degree is a matter of serious concern. This study illustrates the factors that attract students in selecting the business university. The purpose of this study is to investigate the factors which affect the student’s choice for selecting business school in Pakistan. In our research method, we have chosen convenience sampling and distributed 500 questionnaires online. In this study, student choice criterion is divided further into elements of location, fee structure, distance, foreign collaboration and quality education, which students look for before choice the best institute for them, and how it affects overall on students life.

 

 

 

 

CONSUMER BUYING BEHAVIOR ON E COMMERCE THE CASE OF VEGETABLES

 

 

ANUM KHAN 15K-2704

 

(k152704@nu.edu.pk)

MOHAMMAD DANIYAL15K-2719 (k152719@nu.edu.pk)
ABDUL BASIT 15K-2723 (k152723@nu.edu.pk)

 

 

SUPERVISOR

MUHAMMAD AHSAN (muhammad.ahsan@nu.edu.pk)

 

 

E-commerce is growing at an unprecedented rate in Pakistan. However, in the context of vegetables, the growth rate of e-commerce appears to be dormant. This research explores the relationship among consumer buying behavior and variables including price, quality, convenience and trust issues pertaining to e-commerce. The research is quantitative in nature utilizing primary data which is collected using questionnaires and analyzed using Regression as a test to explore the relationship among the aforementioned variables.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BOARD OF DIRECTORS AND FIRM’S PERFORMANCE

 

 

BAKHTAWAR BALOCH 14K-2742

 

(k152742@nu.edu.pk)

ROZINA JAMALI 15K-2710 (k152710@nu.edu.pk)
SHAZIA SAHEB 15K-2716 (k152716@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

This research paper examines the relation between the effectiveness of board of director and the financial performance of the firms that are listed in Pakistan stock exchange (PSX).The purpose of the paper is to highlight the attribute of top management that impacts positively and negatively on the profitability of the firm, and it is also related with the economy of the country. The study used linear regression analysis over the period of five years from 2012 -2017 using the sample of 30 companies of cement and steel sector that are listed on PSX. The impact of board of director on firm performance is analyzed by seven characteristics of board: Non executive director, age of director, female presence in board, CEO duality, qualification of director, board meetings and board size and the performance of firm measured by financial indicators that are ROA, ROE and Leverage.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EFFECTS OF CAPITAL STRUCTURE ON THE PERFORMANCE OF MANUFACTURING FIRMS IN PAKISTAN

 

 

HASAN RAZA 15k-2714

 

(k152714@nu.edu.pk)

ARUBA 15k-2711 (k152711@nu.edu.pk)
ALI ZAIN 15k-2706 (k152706@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

The impact of capital structure is widely argued that whether capital structure has any effect on firm performance or not. The Modigliani and Miller theory of capital structure suggests no relation. On contrary, trade off theory argues that on some extent, financing by the Debt saves tax and benefits organization. The aim of our research is to determine the effects of capital structure on manufacturing companies of Pakistan. The Control variables are short term debts and long term debts; the dependent variable we have used is Return on Equity (ROE). The needed data for the research is collected from the archive, annual reports of companies. The sample contains 45 manufacturing firms listed on Pakistan stock exchange since 2014. The companies are randomly selected from different manufacturing sector i.e. cement, fertilizer, chemical, and automobiles. Correlation method is used to identify the relationship between the Debts and the Return on Equity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IMPACT OF RETAIL LAYOUT, DESIGN AND LOCATION ON CONSUMER BUYING BEHAVIOUR

 

 

RAJ SINGH 15k-2709

 

(k152709@nu.edu.pk)

SARANG MOGHAL 15k-2708 (k152708@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

The purpose of our research is to find the gap of location in retail supermarkets along with other factors that affect retail sale and contribute in retail sales are also considered. Major focus of our research is retail location (i.e. Traffic within store and Parking) and also to focus on retail design and retail layout. This research also identifies the relationship between dependent and independent variables which are buyer buying behavior (Demographics) and Instore traffic respectively. This is the primary research and data will be collected through questionnaire and for analysis SPSS software will be used. In this study samples of 400 respondents will be observed to measure the impact of purposed variables.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EXPLORING STUDENT CHOICE CRITERIA FOR SELECTING BUSINESS SCHOOLS IN KARACHI

 

 

SHAZRA MANZOOR 15k-2702

 

(k152702@nu.edu.pk)

ARFA QAMAR 15k-2712 (k152712@nu.edu.pk)

 

SUPERVISOR

SYED BABAR ALI (babar.ali@nu.edu.pk)

 

This paper investigates the selection criteria for choosing business school in Pakistan. In a student’s life, choice of institution which he/she is going to attend after the FSC/intermediate/A levels is of critical nature. Because it includes so many decisions before getting their admission done in preferred university, in this way many factors influence this choice made by the students. These may include the perceptions of people about universities, parental choice, and different factors such as location, distance, quality of education, faculty, fee structure, campus and facilities, recognition in job market, and alumni strength which students generally develop on their life. It is because every student wants to enroll in good and reputed university which can help them to target their future goals. For students the selection of an institution to obtain a degree is a matter of serious concern. This study illustrates the factors that attract students in selecting the business university. The purpose of this study is to investigate the factors which affect the student’s choice for selecting business school in Pakistan. In our research method, we have chosen convenience sampling and distributed 500 questionnaires online. In this study, student choice criterion is divided further into elements of location, fee structure, distance, foreign collaboration and quality education, which students look for before choice the best institute for them, and how it affects overall on students life.