Zyena Kamran

Academic

Verbatim text

Zyena Kamran
03364867577
Executive Lodges, Phase 3, Bahria Town, Islamabad
LinkedIn:
Education
FAST NUCES Islamabad
BS(CS)
CS, NLP, SMD
Roots IVY International
Math, Chemistry, Physics
Benchmark School System
Math, Physics, Computer Science
Projects
Final Project:
App Feature Recommendation Engine (Python, Flask, React)
Automated app feature recommendation, aiming to provide developers insights for informed decision-making and
competitive development.
Semester Projects:
MusicHub Android Application (PHP, SQL, Java)
Online Music Application inspired from Spotify. Implemented secure Firebase Authentication, and efficiently managed
music files using Firebase Storage and enabled cloud storage.
On-Page and Off-Page SEO Optimization (WordPress, HTML, CSS)
Developed and managed a website using WordPress. Monitored and analyzed user statistics using Google Analytics
for data-driven insights.
Work Experience
C++ Development Intern, MRS Technologies/Technology Spirits
July 2023 - August 2023
Developed an admin portal for a microgreens IoT system to streamline statistics management for administrator.
Worked using QT Core/GUI and QML on Desktop development in C++, AWS IoT Core/ AWS SDK C++ and MQTT to
communicate with Iot Things.
Member of Parallel Computing Networks Lab, FAST University
September 2021 - June 2023
Contributed to diverse projects ranging from Web Development to AI Machine Learning. Documented a Python
library: Urduistics.
Technovation Girls: Challenge Team Lead and Mentor, Technovation
March 2016 - March 2016
Designed an android mobile application for blood donation. Pioneered the concept of E-donor cards and led and
mentored a team of 8 students.
Skills & Tools
Professional Skills:
Problem Solving, Leadership, Communication
Technical Skills:
MERN, Python, Flask, Django, Java, JavaFX, C, C++, C#, SQL, PHP, Hadoop, MPI, OpenCL
Achievements
Gold Medal in Semester 7, Gold Medal in Semester 5, and Silver Medal in Semester 6, FAST NUCES.
Merit Scholarship in O levels and A Levels. Achieved 2 A*s and 1 A in A Levels and secured 7 A*s and
Rector's List 2022, 2023 at FAST NUCES. Dean's Honor List Jan 2021-Present at FAST NUCES.
Activities
Led the Social Media Department for Fast Society of Arts
Member of Fast Female Volleyball Team
Organizer of Devil's Lair Nascon'23
Interests
Baking & Cooking, Sports, Writing
zyenak@gmail.com
https://www.linkedin.com/in/zyena-kamran-970b1127a/
Majors:
A Levels (
)
O Levels (
)
O-Rental (AI-Rental)
ORental aims to revolutionize the rental market by providing a platform that utilizes
advanced automation and intelligence. It serves as a marketplace where individuals can
rent out their belongings, maximizing their utility while fostering a sense of community and
trust among users. ORental aims to be more than just a rental marketplace; it's a platform
that empowers users to monetize their belongings while fostering community bonds and
ensuring a safe and user-friendly ecosystem
Key Features:
Inventory Management: Efficient management of available products for rent.
Wear and Tear Analysis: Analyzing the condition of products to ensure quality.
Cart Management: Streamlined process for managing orders, checkout, and adding to
cart.
AI-based Recommendation and Ranking: Personalized recommendations and user
rankings for enhanced user experience.
Product Management: Comprehensive management of product listings.
Booking System: Booking details management including calendar integration.
Listings Management: Listings for both renters and buyers.
Chat Bot Integration: Assistance and support through a chatbot interface.
Technology Used:
React, Node.js, PostgreSQL, Flask, Langchain,
RAG Model, YOLO  Object Detection, Jupyter
Notebook (Anaconda), Visual Studio Code
Supervisor Name:
Dr. Atif Jilani
Group Members:
Muhammad Daniyal (i20 - 0402), 0331-5546484
Muhammad Mujtaba (i20 - 0649), 0321-1949672
Aina Zainab (i20 - 0900), 0332-9600559
AnkhCheck
AnkhCheck is a cutting-edge eye disease recognition system leveraging deep learning and
image processing techniques. Specifically designed for classifying retinal diseases such as
Diabetic Retinopathy, Myopia, Hypertensive Retinopathy, Glaucoma, and Normal conditions,
AnkhCheck ensures accurate and efficient diagnosis.
The architecture involves statistical preprocessing to achieve a higher accuracy in disease
diagnosis. The system culminates in a user-friendly web application interface, making its
capabilities accessible and convenient for healthcare professionals and patients alike.
Technology Used:
Flask, PyTorch, React, OpenCV
Supervisor Name:
Dr. Labiba Fahad
Group Members:
Muhammad Subhan (i20 - 0873)
Jawad Ahmed (i20 - 0945)
Cloud Fusion
Cloud Fusion has simplified the complex process of cloud provisioning and deployment. We
have created a platform that allows users to effortlessly input their specific cloud
requirements, triggering automated creation of virtual machines on platforms like Amazon
AWS and Microsoft Azure. From this stage forward, we have automated the deployment
procedure for the user based upon their input of desired architecture. This has been
achieved through meticulously designed orchestration scripts, and the system has also
handled intricate network configurations to ensure seamless integration into the client's
environment.
- User signs up to the “Cloud Fusion” portal, logs in with their GitHub account. All
repositories are listed right away. User selects the repository to be deployed, selects the
cloud service (I.E AWS, MS Azure, Digital Ocean, Fusion Cloud- The private cloud we have
set up at FAST NUCES).
-User enters desired configuration, selects from different services, security groups,
databases and requests deployment. A virtual machine with all selected configurations is
created and application is deployed.
-CI/CD Pipeline is automatically created for the deployed application and any changes made
in repository reflect in the deployed application.
-User is notified of the successful deployment of application.
-Docker deployment has also been incorporated as an additional feature for premium user
experience.
Technology Used:
Next.JS, Node.JS, Git, Terraform, Ansible
Supervisor Name:
Dr. Muhammad Aleem
Group Members:
Moeez Ali Mazhar (i20 - 0526)
Salis Bin Salman (i20 - 0489)
Hasan Murad (i20 - 0792)
Feature Quest
Gait recognition refers to the identification of individuals based on features acquired from
their body movement during walking. Despite the recent advances in the feature extraction
process, variations in things such as camera angles, subject pose, occlusions, and clothing
make it a computationally challenging task to extract an optimal set of features from gait
data. Our research aims to contribute to the state-of-the-art with a machine learning
framework for extracting robust gait features from gait data that both reduces the
dimensionality of the gait data and maximizes the class separability. This will help us to
learn what aspects of walking people generally differ and extract those as general gait
features, improving the discriminative power of our feature set. To identify people without
the need for group-specific features is convenient as particular people might not always
provide annotated learning data. Our research aims to explore Mocap datasets especially
as they remove the need for preprocessing, allowing us to focus solely on optimizing the
feature extraction process.
Features Include:
-
Using Machine Learning to devise a method for efficiently recognizing individuals
from their walking patterns.
-
Receiving better evaluation metrics as compared to other 9 implemented methods.
Technology Used:
Java, Python, MATLAB, React, Visual
Studio
Supervisor Name:
Mr. Shoaib Saleem Khattak
Group Members:
Muhammad Abdullah (i202357)
0328-1666698
Maryam Moeed (i202470)
0332-5011489
Muhammad Bilal Badar (i200768)
0331-5432222
SmartLoan
SmartLoan is a pioneering platform designed to transform the traditional loan
management system by leveraging advanced technologies such as artificial intelligence
(AI) and blockchain. SmartLoan aims to simplify and secure the loan application process
for both borrowers and financial institutions. With a focus on transparency, efficiency, and
user-friendly experiences, SmartLoan introduces a dual-view system:
Features
Borrower View: A user-centric interface allowing loan applicants to submit applications,
upload documents securely, and track loan status in real-time.
Intelligent Data Extraction: The OCR feature is equipped with intelligent algorithms
capable of recognizing and categorizing various types of documents and extracting
relevant data fields.
AI-Driven Analysis: Advanced algorithms for automatic financial document analysis,
enhancing the accuracy of credit assessments.
Block chain Transparency: Utilization of block chain technology to ensure secure,
transparent, and tamper-proof loan management processes.
Technology Used:
React.js, JavaSpringBoot, Blockchain,
MYSQL,
Supervisor Name:
Dr. Muhammad Asim
Group Members:
Abdullah Ranjha (i20-0692))
+92 3215660381
Syed Ammar Hussain Shirazi(i20-0409))
+92 312 5367400
Fatima Anwar (i20-2466)
+92 302 2458509
News Bias Detector
The "News Bias Detector" project is an innovative research initiative designed to
tackle the prevalent issue of bias in news reporting concerning major political parties
biased news can have on shaping public opinion and decision-making, this project
seeks to develop a sophisticated solution for identifying and quantifying biases within
news articles. By doing so, it aims to uphold the principles of impartial journalism,
ensuring that the public receives accurate and unbiased information, which is
fundamental for nurturing a well-informed democratic society.
Features include:
•
Bias Detection: Accurately detects and quantifies bias in news about
•
Unbiased Journalism Promotion: Counters biased reporting to foster
integrity in journalism.
•
Research Gap Bridging: Addresses the lack of bias detection research in
•
Innovative Debiasing Tools: Offers tools to correct bias in news, enhancing
credibility.
Technology Used:
Python, React, PyTorch, GitHub,
Flask
Supervisor Name:
Mr. Saad Salman
Group Members:
Sherwan Qadir (i20 - 0689)
Contact No: 0334-7563515
Mohammad Aosaf (p20 - 0619)
Contact No: 0323-5943577
Muhammad Zain Imran (p20 - 0199)
Contact No: 0333-5704446
Historical 3D Scene Generation
Project Overview
The 3D Scene Generation Model operates through a desktop application that accepts
textual prompts from users about specific historical events. Utilizing a keyword-based
approach, the application offers options closely related to the inputted prompt. Upon
selection, the tool employs web scraping techniques to gather detailed information on the
chosen event, organizing the data chronologically. This refined data is then processed
through a series of sophisticated models including ChatGPT for textual object descriptions
and Stable Diffusion for generating corresponding 2D imagery. Our proprietary algorithm
completes the process by transforming these 2D images into detailed 3D models, ready
for exploration in Unreal Engine 5.
Key Features
Interactive Prompt System: Users can easily navigate through historical events with a
keyword-driven interface, enabling quick access to subjects of interest.
Advanced Data Gathering: Leveraging web scraping technologies, the project
meticulously compiles comprehensive details on selected events, ensuring accuracy and
depth.
Multimodal Integration: Incorporates the capabilities of ChatGPT for detailed object
descriptions and Stable Diffusion for image creation, enriching the 3D modeling process.
3D Scene Generation: Utilizes a cutting-edge algorithm to convert 2D images into 3D
models, displayed in Unreal Engine 5 for an immersive experience.
Technology Used:
Python, WikiData, SPARQL, Unreal
Engine
Supervisor Name:
Dr. Naveed Ahmed
Group Members:
Abdullah Chaudhry (i20-2411)
+92 308 5008982
Kamal Ahmed (i20-0749)
+92 333 5157575
Mutharib Ayub (i20–0476)
+92 321 2990921
AdaptiMart
This project is being done in collaboration with Upstart Commerce and the Knowledge
Discovery and Data Science Lab at Fast. The project will encompass the creation of a web-
based e-commerce platform. This platform will consist of all the traditional features of an e-
commerce platform from the perspective of the admin and the user. Our platform will also
incorporate an innovative AI-powered module. This AI-powered module is the demand of
our industry partners.  It will be developed as a separate microservice that can be integrated
with any E-commerce store.
Features Include:
Inventory Management: Tools to monitor and update inventory
Order Management: Features for placing orders and order history.
Product Catalogue: Easy-to-browse categories, product descriptions, reviews, and ratings.
Pricing Module: AI-powered price optimization.
Promotions: Market basket analysis and promotional offers.
Cart Management: Managing and editing shopping carts.
Search and Navigation: Advanced search with filters and sorting options.
Data Visualizations and Statistics: Visualizations and statistical analyses to offer valuable
insights.
Technology Used:
React, Node, MySQL, FAST API, Python
Supervisor Name:
Dr.M. Faisal Cheema
Group Members:
Ateeb Ahmed (i20 - 0550) 03318815667
Saad Khan (i20 - 1826) 03217661868
Zubair Fawad (i20 - 1755) 03002212345
AdCen
Our project aims to establish an innovative web-based centralized admission platform
streamlining the application process for students applying to multiple universities in
features aimed at streamlining the application process and providing invaluable insights for
informed decision-making. For students, our platform offers a unified application form for
multiple universities, university entry test preparation, real-time application updates, and
insights into trending jobs worldwide. Meanwhile, for parents, our platform provides annual
cost estimations covering various expenses associated with university education, including
tuition, food, accommodation, and transport, empowering them to plan their finances
effectively. By seamlessly integrating these features, our project seeks to simplify the
complexities of university admissions while offering comprehensive support for students'
academic and career journeys.
Technology Used:
MERN, Selenium, Beautiful Soup, Python,
JavaScript, Git, Jupyter
Supervisor Name:
Dr. Ahmad Raza Shahid
Co-Supervisor Name:
Dr. Khubaib Amjad
Group Members:
Fiza Ahmad (i20 - 0506)
03090732287
Alishba Nadeem (i20 - 0595)
03365350507
Momina Minahil (i20 - 0740)
03128708495
AI Converse-Sphere
AI-ConverseSphere is a conversational app that is being developed to give a transformative
solution that allows users to upload documents of their choice and engage with a dynamic
conversational avatar by asking questions to it (both audio or textual) and getting responses
(audio or textual) from the avatar in return. The user will first log into the system. After that
the system will allow the user to upload documents in real-time then hey will be presented
with the main menu, from which they can choose between textual input/output or auditory
input/output. and ask questions regarding that document to the avatar. The avatar will
generate the responses to the queries. This innovative platform will offer both textual and
auditory responses, ensuring accessibility for individuals with reading disabilities.
Features include:
-Users can easily upload documents in pdf format.
-The platform features an intelligent conversational avatar that can understand user queries
and provide relevant responses based on the content of the uploaded documents.
-Users have the option to interact with the conversational avatar using both text-based
input/output and auditory input/output.
-The conversational avatar generates responses in real-time.
Technology Used:
Python, LangChain Framework, OpenAI,
Unreal, Eleven Labs
Supervisor Name:
Dr. Adnan Tariq
Group Members:
(03370422281) Malaika Abbasi (i20 -
0502)
(03436996809) Roha Kabir (i20 - 0552)
(03155118648) Uswa Khan (i20 - 0809)
AIJusTech
AIJusTech is a web-based application designed to provide accessible and accurate legal
users instant access to comprehensive legal information, including constitutional laws and
precedential cases.
The platform aims to address the critical need for accessible legal resources online, offering
guidance to individuals, businesses, and legal practitioners alike. In a country where
navigating the intricacies of the legal system can be challenging, AIJusTech serves as a
valuable tool for enhancing legal literacy and facilitating informed decision-making.
Technology Used:
Langchain, Next JS, AWS
Python, Visual Studio, Firebase
Supervisor Name:
Dr. Mehreen Alam
Group Members:
Ehtisham ul Hassan (i20 - 0462)
Cell # 03035184817
Danyal Aziz Memon (i20 - 1811)
Cell # 03461157881
Humaid Ashraf (i20 - 1813)
Cell # 03322098403
APPFIRE - App Feature Identification & Recommendation
Engine
The past decade has witnessed an exponential rise in smartphone users, with over billions
of smartphone owners globally. This surge in mobile device usage has led to intense
competition in the mobile app market, necessitating continuous innovation in app features
to attract and retain users. This project seeks to address this challenge by investigating the
potential of app store data for feature recommendation and prioritization, aiming to enhance
the quality and competitiveness of mobile applications.
Contributions:
● Compiled a dataset of 219 apps and manually annotated app descriptions with 738
features, categorized into 608 functional and 130 non-functional features.
● Developed a manually annotated dataset for feature identification in user reviews,
consisting of 2000 user reviews.
● Proposed a novel approach for feature extraction from both app descriptions and
user reviews by fine-tuning NER Model and applying Semi-supervised learning
techniques.
Technology Used:
Python, Flask, React, MongoDB
Supervisor Name:
Dr. Khubaib Amjad Alam
Group Members:
Ramsha Ali (20I-0839)
Zyena Kamran (20I-0802)
Sabeen Fatima (20I-0505)
Art-Evolve
inclusive online platform for exhibiting and selling artworks. Traditional in-house auctions have
limitations in terms of target audience length and bidding timeframe, constraining both
consumers and artists. This platform seeks to overcome these obstacles via providing
convenience for consumers to discover and bid on artworks and for artists to exhibit their work to
a broader target market.
The number one objectives of the challenge encompass creating a consumer-pleasant auction
platform, implementing standard bidding procedures, permitting customers to bid on artworks
listed via different artists, and custom artwork requests.
The undertaking scope encompasses the development of an internet site with functions for
profile management, auction scheduling, bids placement, custom paintings proposals, image
captioning, sentiment analysis, and recommendation system.
Technology Used:
React, Python, Firebase, Figma, Hugging Face,
MongoDB, JavaScript
Supervisor Name:
Ms. Tajwar Mehmood
Group Members:
Fatima Farooq (20i - 2304)
0346 5001034
Kashfa Farooq (19i - 0421)
0349 6688103
Wasiq Majeed (19i - 0418)
0300 6320882
WheatInsight
WheatInsight is an innovative web application designed to provide methods and strategies
to improve wheat yield by providing advanced features for spike quantification and disease
detection. It leverages image processing techniques and deep learning algorithms to
analyze images of wheat spikes and detect any signs of diseases, aiding farmers in making
informed decisions for crop management.
Features:
-Wheat Spike Quantification: WheatInsight allows users to upload images of wheat spikes,
and through advanced image processing techniques, it provides accurate quantification of
wheat spikes. This feature helps farmers in assessing the yield potential of their crops.
-Wheat Disease Detection: The application utilizes machine learning models trained on
datasets of wheat diseases to analyze images and detect any signs of diseases affecting
the wheat crop. Early detection of diseases enables timely interventions, reducing crop
losses.
-User-friendly Interface: WheatInsight features a user-friendly interface designed to facilitate
easy navigation and seamless interaction. Users can easily upload images, view results,
and interpret analysis reports without any technical expertise.
Technology Used:
Python, Google Colab, Flask, React
4.6, Visual Studio Code, Ultralytics,
Pytorch, Tensorflow.
Supervisor Name:
Dr Labiba Fahad
Group Members:
Rabia Kewan (i20 - 2491)
Rumaisa Ilyas (i20 - 0664)
Fahad Ramzan (i20 - 0443)
Bits per Pixel Steganography
Our Final Year Project (FYP) focuses on developing an advanced image steganography
system that utilizes a combination of neural networks, specifically convolutional neural
networks (CNNs) and a Vector Quantized Variational AutoEncoder (VQ-VAE), to enhance
the hiding capacity of cover images. The project's primary objective is to embed multiple
secret images into a single cover image, ensuring that the steganography process is
imperceptible to the human eye and robust against various forms of detection.
The project involves the creation of a steganographic model capable of processing and
encoding images with high efficiency and discretion. By leveraging the strengths of CNNs
for feature extraction and the VQ-VAE for generating discrete latent space
representations, the system aims to hide information more effectively within cover images.
The application serves the dual purpose of securing sensitive information through
steganography and demonstrating the potential of deep learning in information hiding
techniques.
Notable Features:
- The model significantly increases the amount of information (secret images) that can be
hidden within a single cover image without compromising the image quality.
- Integrates CNNs to efficiently extract and learn features from both cover and secret
images, utilizing these features to better blend the secret images into the cover images.
- Employs a VQ-VAE architecture, which encodes images into a discrete latent space,
enhancing the model's ability to reconstruct high-quality stego images.
- Uses a custom loss function to optimize the balance between the visibility of the stego
images and the fidelity of the embedded secret images, ensuring minimal distortion.
Technology Used:
Python, React, Google Colab, Visual
Studio Code
Supervisor Name:
Ms.Marium Hida
Group Members:
Tayyab Imtiaz (20i-2433)
Kashan Altaf (20i-0547)
Umair Amjad (20i-0960)
Book2Life
Book2Life presents a groundbreaking web application for integrating engaging illustrations into PDF
books, ending the dull reading experience. Through cutting-edge artificial intelligence (AI)
technology, authors and publishers can effortlessly elevate their narratives without the need for
extensive illustrator investments.
Our mission is to develop a web platform accessible to content creators of all technical backgrounds,
facilitating PDF book illustration for all. We offer a diverse range of illustration styles, from
Cartoonish to Realistic and Line Art/Sketch, catering to authors' diverse preferences and artistic
visions.
Our approach encompasses:
• Creation of user-friendly dashboards for streamlined content management and
customization.
• Utilization of cutting-edge Language Model (LLM) technology for text extraction and prompt
generation, enhancing character and scene depiction accuracy.
• Integration of a stable diffusion model fine-tuned with extensive online image datasets,
ensuring the fidelity of generated illustrations.
• Implementation of mechanisms to maintain character consistency throughout the narrative,
guaranteeing a cohesive storytelling experience.
Technology Used:
Python, Stable Diffusion, Llama 2, MongoDB,
ReactJs, NodeJs, Google Colab
Supervisor Name:
Dr. Atif Jilani
Group Members:
Abdul Wahab (20i-0465) 0318 5062949
Ahsan Rasheed (20i-04740) 0348 8539914
Muhammad Huzaifa (20i-2473) 0300 5990259
Caption Craft
Our team has developed a mobile application aimed at revolutionizing the process of
crafting product descriptions for retailers. Leveraging advanced captioning models and
large language models, our application streamlines the time-consuming task of generating
compelling marketing product descriptions. Furthermore, we've incorporated a unique
feature that empowers users to append additional specific features to enhance the
generated descriptions, ensuring a tailored
Features include:
● Auto-filling of Meta Information -
● Description Generation
● User Customization
Technology Used:
Python, Hugging Face, GCP, Fluter
Supervisor Name:
Dr. Asif Naeem
Group Members:
Muhsin Raza (i20-1759) 0309-0795658
Shahwaiz Memon (i20 - 0681) 0331-
3468427
Eisha Rehan (i20-0539): 0310-8326844
Cardiac Anomaly Detection
We have developed an advanced tool for the initial assessment of cardiac anomalies by
leveraging recent innovations in signal processing, machine learning, and medical
research. Our project aims to bridge the gap in reducing model complexity while
enhancing accuracy in end-to-end cardiac anomaly detection, specifically utilizing
Electrocardiogram (ECG) and Phonocardiogram (PCG) data alongside AI.
Features:
Data Input: Users can input ECG, PCG, or both for analysis.
Visualization: Real-time display of raw and filtered data.
Anomaly Detection: Instantaneous analysis for anomaly presence.
Research Section: Insights into methodology and findings.
Team Introduction: Meet the project team.
Technology Used:
Pandas, Flask, NumPy, React JS,
TensorFlow
Supervisor Name:
Dr. Uzair Iqbal
Group Members:
Mahnoor Akhtar (i20 - 0635)
Ubaidullah Javed (i20 - 0660)
Rafay Zubair Gill (i20 - 2445)
CarMeetUp (Wheels Unite)
Our app is designed for car enthusiasts to plan and manage car shows and car meetups.
We have provided them with a dedicated platform for all their car ventures. Our project
offers a diverse range of features and functionalities to meet their needs. Among its core
offerings are user profiles, event management tools, communication features, and a
marketplace for buying and selling cars and related items.
Features include:
- 1. User-Friendly Interface
2. Event Creation and Participation
3. Location-Based Meet-ups
4. Car Showcase
6. Notification System
9. In-App Messaging
Technology Used:
Kotlin, Firebase, C#, Unity , .Net , Visual
Studio, Github, MySql
Supervisor Name:
Mr. Saad Salman
Group Members:
Ahmad Maqbool (20i-0702)(03010506152)
Aizaz Ahmed (20i-0406)(03239595626)
Waleed Ahmad (20i-0858)(03017581839)
CoPlan AR – Collaborative AR Application for Floor Planning
Traditional floor planning methods often lack real-time collaboration, limited by physical
models, 2D drawings, and disconnected communication. CoPlan AR's mission is to
revolutionize floor planning in the architecture and design domain by using the power of
collaborative augmented reality. It provides an AR-based mobile application that enables
collocated users to collaborate by overlaying digital models onto the real-world
environment and editing them in real-time. Users create or join a collaborative session,
select from a list of relevant floor planning assets, place, orient, scale and move them as
they see fit. Changes incurred by these adjustments reflect in the interfaces of all other
users that are part of the same session with realistic precision in real-time.
The main technical features of CoPlan AR include:
• Multiplane Detection
• Object Placement, Movement, Rotation, Scaling
• Anchor/Transform Management
• User Management
• Collaboration & Network Synchronization
GitHub
https://github.com/aybeedee/coplan-ar
Website
https://www.coplan-ar.com
Release
https://github.com/aybeedee/coplan-
ar/releases/tag/Latest
Technologies Used
Unity, C#, ARCore, Firebase, Niantic
Lightship ARDK, Netcode, React,
Next.js, Vercel, Figma, GitHub
Supervisor Name
Dr. Adnan Tariq
Group Members
Abdullah Umer (i20-0528)
+92 3325727893
Ramish Afaq (f20-0433)
+92 3374922661
Muhammad bin Awais (i19-0431)
+92 3329415030
CropCartel
We are making an innovative mobile application poised to revolutionize the agricultural
landscape. Our mission is to empower the agricultural community through technology,
facilitating efficient crop transactions. This mobile-driven solution will seamlessly connect
traders and buyers with each other and streamlining the entire crop trading process. Beyond
facilitating transactions, our application boasts cutting-edge AI features, including crop
recommendation, disease detection and Yield Prediction.
Features include:
-
Marketplace Functionality
-
Messaging and Communication
-
Inventory Management
-
Bidding System
-
Disease Detection in Crops
-
Yield Prediction in the Crops
-
Crop Recommendation System
-
Review and Rating System
-
Dual Language (English & Urdu)
Technology Used:
Pytorch, OpenCV, Flutter, Android
Studio, Flask, Firebase
Supervisor Name:
Mr. Saad Salman
Group Members:
Atif Munir (i19-0600)
03086583548
Ahmad Amjad (i20-0420)
03405922265
Shaheer Abdullah (i20-1861)
03079540000
CyberFront-AI
CyberFront-AI is an innovative project aimed at simulating a red team vs blue team
scenario within a Docker environment. The project encompasses a suite of scripts
designed to exploit vulnerabilities and misconfigurations in Docker containers and images
(red team), as well as scripts to patch and mitigate these vulnerabilities (blue team).
Additionally, a machine learning model has been developed to predict potential attacks on
Docker containers based on benchmark security metrics. To facilitate easier access to the
extensive benchmark documentation, a chatbot has been implemented to provide quick
and accurate answers to queries.
Features:
1. Red Team Scripts: Targeting various vulnerabilities and misconfigurations within
Docker environments.
2. Blue Team Scripts: Developed to prevent exploitation by malicious attacks.
3. AI Prediction Model: Trained to predict the likelihood of specific attacks targeting
Docker containers.
4. Documentation Chatbot: A chatbot has been implemented to provide fast and
accurate answers based on the extensive benchmark documentation, simplifying
access for users.
Technology Used:
Python, Docker, Mern Stack, Hugging
Face, Git
Supervisor Name:
Dr Qaisar Shafi
Group Members:
Talha Atif (20i-0486)
03368240992
Shahbano Baber (20i-0607)
03149797782
Usman Nadeem (20i-0551)
03175070955
Deepfake Detect
Deepfakes, sophisticated video and image manipulations enabled by advanced artificial
intelligence (AI), present a growing challenge in digital media integrity. They can be used
maliciously to spread misinformation, commit fraud, or tarnish reputations, with potentially
severe consequences for individuals. Our project aims to contribute to the battle against
digital forgery by replicating and enhancing the Spatial-Frequency Dynamic Graph (SFDG)
method proposed by Yuan Wang et al. for deepfake detection. Our objective is to develop a
more nuanced and adaptive approach that can effectively identify a wider range of
deepfakes, including those not yet encountered during training. By analyzing the
relationships between spatial and frequency domain features through dynamic graph
learning,
we
seek
to
improve
the
model's
accuracy
and
generalizability.
Features include:
•
Adaptive Detection: Leverages dynamic graph learning to adaptively identify and
analyze deepfakes, improving detection accuracy across varied forgery types.
•
Spatial-Frequency Analysis: Integrates spatial and frequency domain features for
comprehensive analysis, capturing subtle manipulation cues invisible to the naked
eye.
•
Enhanced Data Processing: Utilizes advanced data augmentation to enrich training
datasets, simulating a broader range of forgery scenarios for robust model training.
•
User-friendly Demonstration: Features a prototype application developed with
Streamlit, showcasing the model's detection capabilities in an interactive manner.
Technology Used:
Python, TensorFlow, NumPy, Pandas,
Matplotlib, Seaborn, OpenCV, Dlib,
Streamlit, Google Colab
Supervisor Name:
Ms. Marium Hida
Group Members:
Saad Ahmad (i20-0508)   03331004981
Muhammad Usman (i20-0507)
03317748119
DubLingo
Project Scope: DubLingo is an automated dubbing system designed to seamlessly dubbed
Urdu drama videos into Arabic, enhancing accessibility for Arabic viewers.
•
Audio Processing: Utilizing advanced techniques, we will separate vocals from
background music in the videos.
•
Text Extraction: The extracted vocals will be processed to achieve accurate Urdu
text transcriptions with timestamps and speaker diarization.
•
Voice Synthesis: Employing AI-powered voice synthesis, we will generate high-
quality Arabic voice-overs that mimic the original voice characteristics.
Impact and Goals:
•
Bridge the language gap: By automating the dubbing process, DubLingo seeks to
offer Arabic audiences a wider range of high-quality Urdu content, removing linguistic
barriers and enriching their viewing experience.
Technology Used:
Nextjs, Flask, Visual Studio, Google Colab
Supervisor Name:
Dr. Mehreen Alam
Group Members:
Zahid Imran 20I-0469
+92 323 6512178
Ali Tajir  20I-0512
+92 300 4564477
Muhammad Shezad 20I-1756
+92 302 0104553
EaseAssist
EaseAssist is a centralized platform for easy access of scholarships for students and
management of those scholarships for universities. The application has mainly three views:
Student View, University View, and Mentor View. Students can access all scholarships along
with certain recommended ones. Students can upload their documents and they can also
apply for their preferred scholarships in one click later on. On the other hand, universities can
view valuable insights on the scholarships they have posted.  Universities also get the list of
shortlisted students with whom they can schedule meetings through our platform for further
processing. Mentors are nominated by universities and through their mentor view, They can
manage their profiles and guide the reached out students through chat. Following are some
prominent features of our application:
● One click apply on scholarships.
● Personalized recommendation model for recommending scholarships.
● Shortlisting and automated scheduling of meetings for universities.
Technology Used:
MERN Stack, Python, Tensorflow, Flask,
Scrapy, AWS, GitHub, Postman
Supervisor Name:
Ms. Maheen Arshad
Group Members:
Muhammad Shayan Khan (i19 - 0743)
0318-0985447
Maryam Zafar (i20-1874)
0312-5246630
Alina Asim (i20–0628)
0300-5386822
Entekhaab
A Blockchain-Based Voting & Management system, built following a Microservices
architecture, encompasses four key user roles: Voter, Candidate, Party, and Admin. Each
user type has unique features tailored to their roles in the electoral process.
The Admin is authorized to create and oversee elections, as well as manage user accounts.
Parties can enlist or remove members and grant election participation privileges to select
members. Candidates, upon approval from both their respective party and the admin, are
eligible to participate in elections. Voters play a pivotal role by casting their votes and can
even switch their voting constituency if necessary. Additionally, voters can anonymously
verify their votes on the blockchain, ensuring complete vote privacy.
Users can also access the Results window to analyze past elections using graphs. Parties,
Candidates, and Voters can refine their strategies, understand trends, and make informed
decisions based on historical data.
Technology Used:
• MongoDB, Express.js, React.js,
Node.js
• Solidity, Polygon
• Docker, AWS
Supervisor Name:
Mr. Muhammad Abdullah Abid
Co-Supervisor Name:
Mr. Zaheer ul Hussain Sani
Group Members:
Abdul Manan  |  i20-2448  |  03355100705
Bilal Khan  |  i20-0542  |  03471623073
Shoaib Ali  |  i20-0548  |  03495102474
EquiSpeak
Our project aims to develop a cutting-edge mobile application that revolutionizes
communication for individuals with hearing impairments. By leveraging advanced
technologies, such as machine learning and real-time processing, the app will enable
seamless interaction through optimized gesture recognition. Alongside this core feature,
the app will offer voice-to-text and text-to-voice conversion capabilities, digital flip cards for
alternative communication, and a user-friendly interface designed for accessibility.
Features include:
Sign Language Gesture Recognition: The app will feature state-of-the-art algorithms for
accurately recognizing and interpreting a wide range of sign language gestures in real-
time.
Voice-to-Text Conversion: Users will have the ability to speak into the app, which will
transcribe their spoken words into text in real-time, enabling smooth communication with
both hearing and non-hearing individuals.
Text-to-Voice Conversion: Conversely, the app will also support text-to-voice
conversion, allowing users to type messages that are then synthesized into natural-
sounding speech, facilitating communication with non-text-based individuals.
Technology Used:
Flutter, MediaPipe, TensorFlow,
Python, OpenCV, Figma
Supervisor Name:
Dr. Atif Jillani
Group Members:
Ammar Altaf (20i-0430)
(+923155109294)
Aashan Javed (20i-0491)
(+923361665553)
Shameer Abdullah (20i-0540)
(+923345555785)
E-Recruitment
The E-Recruitment project represents a significant advancement in the recruitment process,
aiming to enhance efficiency and accuracy through modern technologies. The project
begins by collecting CVs/resumes from job applicants within the specified timeframe,
securely storing them in a database. Once the application deadline is reached, the system
extracts this dataset for further processing. Employing sophisticated Natural Language
Processing (NLP) and Artificial Intelligence (AI) techniques, the system parses the
CVs/resumes to extract vital information such as candidates' education, skills, and projects.
This process, although time-consuming due to diverse CV formats, ensures comprehensive
data extraction.
Moreover, the project integrates with LinkedIn and GitHub APIs to extract candidates' profile
links, subsequently retrieving relevant insights from these platforms. The gathered
information from CVs, LinkedIn, and GitHub profiles is then stored in the database, forming
detailed candidate profiles. Recruiters benefit from a React-based web dashboard that
offers user-friendly access to extensive candidate information. They can review candidates'
education, skills, projects, and LinkedIn/GitHub contributions effortlessly. Additionally,
candidates have a dedicated login to apply for jobs, monitor their application status, and
access personalized dashboard features.
Key features of the E-Recruitment project include automated data collection, advanced NLP
and AI techniques for data parsing, seamless integration with LinkedIn and GitHub, secure
database storage, comprehensive candidate profiles, an intuitive web dashboard, and a
dual login system for recruiters and candidates. By combining technological innovation with
user-centric design, the project revolutionizes recruitment practices, improving the overall
efficiency, accuracy, and effectiveness of the recruitment lifecycle.
Technology Used:
React, Next, AWS, Mongo-DB,
, Collab, TensorFlow
Supervisor Name:
Mr. Shoaib Mehboob
Group Members:
Muhammad Taha(i200485)
Contact:0337-6006066
Huzaifa Bilal (i200568)
Contact:0317-3371929
Haris Mehmood(i200902)
Contact:0316-5492388
Job Connect
The "Job Connect" project is a transformative platform aimed to revolutionize the traditional
hiring and skill assessment landscape. Job Connect empowers job seekers and employers
by facilitating precise skill assessment and alignment. Through intelligent CV parsing and
skill extraction, candidates can effectively showcase their capabilities, while employers can
pinpoint the ideal candidates with precision. The platform dynamically generates tailored
quizzes for candidates, evaluating their skills accurately and fairly.
Features include:
• User authentication and CV parsing
• Skill extraction and job description matching
• Candidate evaluation using questions
•     Evaluation and candidate ranking
Technology Used:
Python, MongoDB, React js, Flask, Node
js, GitHub, Visual Studio
Supervisor Name:
Dr. Kashif Munir
Group Members:
Shayaan Hasnain (i20-0647)
03359412740
Manal Zehra (i20 - 0828)
03095309880
Muhammad Dayyan (i20 - 0655)
03042557799
CarRozgaar
CarRozgaar is an innovative platform designed to bridge the gap between car owners and
advertisers, providing a unique marketplace for mobile advertising. The application has two
primary interfaces: Car Owner View and Advertiser View.
Car Owner View: This interface is tailored for car owners, offering them a user-friendly
platform to browse and register for advertising campaigns. Features include:
•
Marketplace Access: Car owners can explore a wide range of advertisements
posted by various advertisers, along with the specified rates.
•
Registration for Ads: Users can easily register their interest in specific campaigns
and agree to place the advertiser's sticker on their vehicle.
•
Earnings Tracker: A dashboard for car owners to track their earnings, with a
detailed breakdown of basic revenue and bonuses earned by driving through
designated hotspots.
•
3D Preview: An innovative feature that allows car owners to view a 3D model of their
car with the advertisement applied, providing a realistic preview of how it will look.
Advertiser View: Designed for advertisers, this interface provides the tools needed to
create, manage, and monitor advertising campaigns. Features include:
•
Ad Posting: Advertisers can post new advertising campaigns and specify the rates.
•
Campaign Management: Tools for managing live campaigns, updating rates, and
monitoring engagement and reach.
Technology Used:
Flutter, Unity, Blender, React,
Next JS, Firebase, Figma
Supervisor Name:
Mr. Owais Idrees
Group Members:
Ahmad Abdullah (i20 - 1773) (0333
9978339)
Furqan Nasir (i20 - 0413) (0335 7585343)
Uzair Ahmed (i20 - 1751) (0332 1120112)
Student Feedback Analyzer
The Student Feedback Analyzer revolutionizes educational feedback processing. It
analyzes PDFs of student reviews on instructor quality and course aspects, generating
actionable insights. Instructors receive email notifications post-account creation and after
semester feedback analysis. The database tracks instructor progress by subject and
category, aiding informed decision-making. Constructive criticism comments are also
highlighted for instructors to respond to via the system.
• Feedback uploading and data extraction: Seamlessly upload feedback PDFs and
extract relevant data for analysis.
• Sentimental Analysis: Analyze sentiment in feedback to derive actionable insights.
• Role-based actions and interactions: Customize user roles and interactions based
on organizational hierarchy and responsibilities.
• Notification and Communication: Receive timely notifications and facilitate
communication between instructors and management authorities.
• Data Visualization and Analysis: Visualize feedback data for comprehensive
analysis and decision-making.
• Submission of Response: Provide response to highlighted feedback points to foster
dialogue and improvement.
Technology Used:
Django, React, MySQL
Supervisor Name:
Dr. Kashif Munir
Group Members:
Areeba Sattar  (i20  0634)
[+923235544520]
Maria Saeed    (i20 - 0836)
[+923365282377]
Afnan Azhar    (i20 –
0908)[+923318206929]
FigForge – Design to Frontend Code Converter
Description:
Too much time, resources and manpower is spent converting Figma designs to code. This
is why we’ve automated the end-to-end process so that you go from a Figma design file to
accurate React/Next.JS code deployed onto your GitHub profile. This System benefits the
designers/ frontend developers by saving hours of their time by giving them working
frontend code straight from their Figma wireframes.
Architecture:
The system incorporates 4 Main components acting as individual microservices. Firstly,
the Figma Plugin which runs as a direct assistant to users on Figma (React/TS). Secondly,
the code engine, used to convert design data to front-end code (Python/Django). Thirdly,
the Web App to manage user’s projects and generated code (NEXT.JS). Finally, the
Node.JS server to automate deployment.
Features:
-
Managing of Projects
-
Accurate Design to Code
-
Dynamic and intractable components
-
Responsive Designs for Mobile/Tablet/Desktop Screens
-
Customized User Routing
-
Variables and State Management
Technology Used:
React.JS, Node.JS,
Python/Django, PostgreSQL,
Supabase, Figma Developers
Plugin, GitHub, Docker
Supervisor Name:
Dr. Atif Aftab Ahmed Jilani
Group Members:
Mahad Ahmed (i20 - 0426)
+92 322 8001177
Areeb Sajjad (i20 - 0904)
+92 304 1871128
Munaf Ul Hassan (i20 - 0891)
+92 315 7600100
Guardian
With Guardian, you can input your location and day of the week to instantly receive a
detailed crime graph showing the probability of different types of crimes occurring in that
area and time. Whether you're planning a night out or exploring a new neighborhood,
Guardian empowers you with valuable insights to make informed decisions about your
safety.
Features include:
1. Crime Graph Insights: Guardian provides a personalized crime graph based on your
entered location, offering insights into the types and probabilities of crimes in your area.
2. Instant Police Assistance:Simply tap to call the nearest police stations based on your
current location, ensuring swift response in case of any emergency or suspicious activity.
3. Emergency Power Button: In critical situations, every second counts. Guardians
Emergency Power Button immediately notifies your pre-set emergency contacts with just a
press.
4. Voice-Activated Help:With Guardian's Voice Help feature, calling for assistance is as
easy as speaking out. Activate this feature, and both your emergency contacts and nearby
police stations will be contacted promptly, providing you with essential aid and support.
Technology Used:
Python, Flutter, Android Studio, Visual
Studio,Deep Learning
Supervisor Name:
Miss Hina Binte Haq
Group Members:
Mayhan Hazara (i20 - 1804)
Momina Hyat (i20 - 1840)
Mayhan :03482346273
Momina:03161503022
Academic Chatbot
H-AI-re project aims to address the time-consuming and biased nature of traditional
interview processes by implementing an automated interview analysis tool that uses
knowledge graph technology to create context-sensitive interview questions. The project's
goal is to streamline the interview, eliminate unconscious bias, boost the effectiveness of
the hiring process, and thus support the recruiter's decision-making. Current online
evaluation tools available in market lack a human like connection. We focus on bridging the
gap between man and machine to have a more natural interview environment by generating
context-based questions and making the interview more conversational.
Highlighting Features:
-
Technical interview specific Video and Audio Analysis for example, confidence,
attention structured answers, Fluency.
-
Technical correctness evaluations of candidate’s answer.
-
Semantic and contextual validity of candidate’s answer.
-
Generating context aware questions based on candidates’ technical capabilities and
job requirements.
-
Real Time evaluation of candidates answers to adjust difficulty level of follow up
questions form the candidate
Technology Used:
Python, React, Flask, Node js, SparQL,
GraphDB, Firebase, Github
Supervisor Name:
Dr. Faisal Cheema
Group Members:
Muhammad Hanzalah (i20- 0956)
+92 301 5301999
Hadiya Farooq (i20- 0579)
+92 334 5265566
Meerub Shami (i20- 0558)
+92 332 3306667
Game Based Interventions for Individuals with ADHD
In the game Heed Impact, GBIs (Gameplay-Based Interventions) play a pivotal role in
the ADHD treatment narrative. These dynamic tools mirror real-life strategies used in
ADHD treatment, including therapy, medication, and mindfulness. Through interactive
storytelling, players learn to harness GBIs to conquer challenges, build focus, and develop
crucial life skills. Beyond gameplay, Heed Impact fosters empathy and education
empowering players to actively engage in the ADHD treatment journey. It's more than
just a game; it's a transformative experience that promotes understanding, resilience and
positive change for both players and those affected by ADHD.
Features include:
- Interactive Gameplay.
- Interactive environment.
- Quests.
- Character Animation.
- Movement States.
Technology Used:
C#, Unity, Blender, .Net framework
4.6, Visual Studio
Supervisor Name:
Dr. Khubaib Amjad Alam
Group Members:
Asad Khalid (i19 - 0522)
Afsheen Ahmed (i19 - 0444)
Phone Numbers:
+92-3035213160 (Asad Khalid)
+92-3185162727 (Afsheen Ahmed)
Herd Help
Embark on an extraordinary journey into the future of farming with Herd Help! Our state-
empowering them to seamlessly engage with an intelligent Chabot using voice or text
prompts in Urdu. Bid farewell to language barriers as Herd Help forges a vital link between
farmers and critical livestock care, addressing everything from health queries to strategic
business insights.This groundbreaking final year project is more than just an app, it's a
visionary force reshaping the landscape of livestock management. With its user-friendly
interface and cutting-edge AI technology, Herd Help leverages customized Large
Language Models, fine-tuned with real-time data and RAG technology, to provide
lightning-fast responses, tailored guidance, and a treasure trove of knowledge right at your
fingertips. Join the revolution today and be a part of driving forward the thriving livestock
Technology Used:
Flask, TensorFlow, LLM, Gemma by
Google, My SQL, Scrapy, Whisper
Supervisor Name:
Dr. Mehreen Alam
Group Members:
Muhammad Usman (i20-0937)
Contact: 0306 6001567
Ahmad Munir (i20-0977)
Contact: 0306 7235224
Ahmed (i20-1893)
Contact: 0315 8899065
HerHealth: A Women's Health App
Our app, designed with the empowerment and health of women in mind, integrates a LLM
with Retrieval-Augmented Generation (RAG) for comprehensive coverage across various
health domains—Mental Health, Menstrual Health, Cancer, Nutrition, and General Health.
It serves as a holistic platform where users can seek personalized advice, insights, and
information in natural language, making health information more accessible and
understandable.To foster a supportive community, the app includes interactive forums
where users can engage in discussions, create and follow new forums, connect with
others, and express their views through likes and dislikes. This feature not only
encourages the sharing of experiences and advice but also cultivates a space for empathy
and support among women navigating similar health journeys.The app is divided into two
interfaces: User View and Admin View. The User View is crafted for individuals seeking
information or advice and looking to engage with the community. Here, they can easily
navigate through a wealth of health-related information, post questions, and interact within
forums on topics of interest.The Admin View, on the other hand, is geared towards
maintaining the integrity and utility of the platform. It features real-time analytics that track
each query made to the chatbot, categorizing them for insights into the most sought-after
information.
Technology Used:
Android Studio, TensorFlow, Firebase,
LangChain, Transformers, HuggingFace,
Python
Supervisor Name:
Mr. Saad Salman
Group Members:
Murtaza Haider (i20-0822) 0355 4910352
Abdullah Tariq Gill (i20-0503) 0304 5252400
Eesha Shafqat (i20-0707) 0336 1556619
ImageNary
Imagenary aims to develop an innovative multimedia AI system that seamlessly combines
text-to-image synthesis, speech translation, and lip synchronization to create virtual avatars.
At first animated characters would be generated based on the user-provided textual prompt.
Then the user would provide text to generate speech and then finally lip movements would
be added to the generated image creating an animated video, which would compromise
whatever character the user created speaking according to the provided texts. Moreover,
users can also talk to the avatar as a friendly chatbot. They will speak whatever question
they want to ask which will be recorded by the microphone and then the avatar will talk back
in real time. The project focuses on delivering a coherent and natural user experience while
addressing technical challenges inherent in such advanced AI systems. All this will be
integrated in a user-friendly web application.
Features Include:
Text-to-Image Synthesis to create a picture.
Speech Generation on the given textual prompts by the user
Lip Synchronization with the speech translated on the avatar image generated to create a
video.
Real Time interactive Chatbot with whatever character that has been created.
Technology Used:
Python, React, Firebase, Docker, Visual Studio
Supervisor Name:
Dr. Ahmed Raza Shahid
Group Members:
Safa Khan (20i-0407)
Cell # 03319507779
Adeel Afzaal (20i-0487)
Cell # 03180617787
Saleem Raza (20i-0577)
Cell # 03321737370
Automated UI/UX Usability Evaluation Tool
Our project aims to develop an automated methodology based on the 10 Nielsen Principles that
will identify usability issues or problems that a website contains which is crucial in allowing
developers to enhance the user experience and overall usability of the website. By employing
object detection methods and machine learning algorithms, we will generate a comprehensive
report highlighting potential improvements. The models used will not only be detecting binary
features, whether they are present or not but will also be assessing dynamic features in terms of
navigational flow and flow of sequence. A score will be generated at the end of the report which
will determine the “goodness” of a website. Static Evaluation: Automatically extracts URLs and
parses frontend code to identify existing features. Dynamic Evaluation: Provides an interactive,
guided experience using image-based input. Maps the identified features onto Nielsen's
heuristics, providing an overall usability score and separate scores for each heuristic.
Technology Used:
Python, Javascript, Html, CSS , Flask, Visual
Studio, TensorFlow, Pytorch
Supervisor Name:
Dr. Khubaib Amjad Alam
Group Members:
Fatima Mustafa (i20 - 0564)
Alishba Asif (i20 - 0582)
Waleed Adnan (i20 - 0897)
Intellect Slide
Intellect Slide is being made, whilst keeping the perspective of students and instructors
personals. There is one view of this web app – Instructors/Students View – the former is for
the students to use the features provided by our web app to upload pptx files and present
the lecture seamlessly using the advanced voice recognition techniques which includes
speech
recognition,
natural
language
processing
and
emphasis
detection.
Students/Instructors will present the lecture using our web app, our app will automatically
detect the emphasized key-points of the lecture slides, lecture slides can also be navigated
easily using the voice commands of ‘next slide’ and ‘previous slide’.
- Speech to Text Conversion: Convert teacher's spoken words into written text to near real-
time.
- Text Analysis and Content Relevance: Process the transcribed text to identify keywords
and   relevant phrases.
- Slide content extraction and parsing: Extract title, textual content and bullet points from
PowerPoint slides (pptx).
- Dynamic Highlighting and Emphasis Tagging: Relevant text and bullet points should be
highlighted on the slides as the teacher speaks.
Technology Used:
React, Flask, Git, Neon Cloud DB, VS
Code, PostgreSQL, Figma
Supervisor Name:
Dr. Arshad Islam
Group Members:
Abdur Raheem Qureshi (20I-0917)
0336 5974759
Ahmad Yar (20I-0850)
0316 1765528
Muhammad Sheharyar (20I-2403))
0334 0738200
IntelliStudy
IntelliStudy is a state-of-the-art mobile application designed to streamline the academic
lives of university students. By leveraging advanced technologies such as artificial
intelligence and machine learning, IntelliStudy offers personalized study schedules,
accurate grade predictions, and intelligent assignment planning. The app's intuitive
interface and comprehensive features aim to enhance time management and academic
performance, making it an essential tool for students navigating the complexities of
university life.
Key Features:
1. User Authentication and Account Management: Secure login and personalized
account settings.
2. Study Schedule Generation: AI-driven scheduling for daily, weekly, or monthly study
plans.
3. Assignment Complexity Assessment: Tailored study plans based on the complexity
of assignments.
4. Grade Prediction and Analysis: Predictive analytics for course grades based on input
marks.
5. Database Management and API Integration: Robust backend support for seamless
data synchronization.
Technology Used:
Python, Flutter, Flask, Github, MySQL,
Visual Studio
Supervisor Name:
Dr. Faisal Cheema
Group Members:
Maheer Arshad (20L-1221) 0311-0545060
Yasin Jodat (20i-2416) 0312-4690800
Fatima Zubeda (20i-0450) 0343-8804349
Interview Ready
InterviewReady is a mobile app designed to enhance interview skills by simulating realistic
interview environments. It does this by creating a realistic practice environment. The idea
is to help users match their academic knowledge with the skills needed to do well in
interviews. The app offers personalized practice sessions based on the user's resume and
job application. Users can set the interview difficulty level to match their comfort and skill
level. Here's what the app does:
1. Creates custom questions based on the user's information.
2. Employs advanced deep learning models for:
• Audio analysis to evaluate the user's verbal responses, focusing on clarity, confidence,
and relevance.
• Video analysis to assess non-verbal cues such as body language and facial
expressions, contributing to a comprehensive understanding of presentation skills.
3. Provides detailed feedback and progress tracking, highlighting areas of improvement
and demonstrating advancement over time.
InterviewReady is made with Flutter, which works smoothly on different devices. It aims to
make users more confident and give them useful advice to improve their interview skills,
helping to close the gap between what they know and how they perform in interviews.
Technology Used:
Flutter, Supabase, Postgresql, Python,
Hume, Hugging Face, Tensorflow
Supervisor Name:
Dr. Akhtar Jamil
Group Members:
Hamza Imran | i20 - 0437 | 03331313667
Harmain Nasim | i20 - 0883 | 03205030970
Danial Bin Abdullah | i20 - 0705
|03310202649
JanabHazir
HazirJanab stands at the forefront of digital innovation providing local services (mechanic,
electrician, plumber, and carpenter) with its two applications and an integrated admin panel,
revolutionizing the way users access services and e-commerce. The user app facilitates not
only immediate service bookings and inquiries but also hosts an e-market for purchasing
necessary parts and products. Parallelly, the vendor app empowers service providers and
vendors with tools to manage listings, interact with customers, and handle orders. Bridging
these two functionalities, the admin panel serves as a comprehensive management tool,
overseeing operations, mediating between users and vendors, and ensuring service quality
and marketplace integrity.
Key Features:
Specialized Apps: Separate apps for users and vendors, plus an admin panel for
management.
Instant Booking: Easy booking and inquiries for immediate and future services.
E-Market: Integrated marketplace for purchasing service-related parts and products.
Vendor Tools: Vendor app for managing offerings and engaging with customers.
Admin Oversight: Admin panel for platform control, quality assurance, and dispute
resolution.
Technology Used:
Android Studio, MERN, Java, Google
Maps, MySQL, GitHub, Figma, Python,
ABSA
Supervisor Name:
Mr. Bilal Khalid Dar
Group Members:
Mohammad Abdullah (i20-0933)-
03488238344
Faizan Ahmed (i20-0546)-03319409717
Mujeeb Ahmed (i20-0611)-03162535057
Kumon: Kubernetes Performance Monitoring Tool
This project aims to develop the Kumon Kubernetes Performance Monitoring Tool, an
essential solution for enhancing the efficiency and of monitoring Kubernetes environments.
The current monitoring tools like APMs and Sidecars often result in operational challenges
and suboptimal resource utilization. Our tool seeks to revolutionize Kubernetes
performance monitoring, enabling operators and developers to manage clusters and
optimize resource allocation through the integration of advanced technologies such as
Cilium, Hubble, Grafana, and eBPF. We aim to provide real-time insights into resource
usage, network traffic, and cluster health metrics, empowering organizations to achieve
greater scalability and cost-effectiveness in their Kubernetes deployments.
Today, the Kubernetes container orchestrator is an indispensable tool in the context of
developing modern applications, supporting crucial workloads in various businesses.
Technology Used:
Kubernetes, eBPF, Cilium, Grafana,
Hubble, and Prometheus.
Supervisor Name:
Dr. Arshad Islam
Group Members:
Abdullah Waqar (i20 - 0615)
Ahsan Iqbal (i20 - 1814)
Rida Kamal (i20 - 2496)
making it accessible to both laymen and legal professionals. Powered by a state-of-the-art
Language Learning Model with over 7 billion parameters, it is meticulously trained on
SECP Acts, ensuring precise and relevant legal advice. With a user-friendly interface
developed in Angular and a robust backend in Python and Flask, LegalBot combines
technical sophistication with ease of use. Its advanced AI, leveraging a vector database
and RAG technology, provides clear, authoritative answers to complex legal queries,
bridging the gap between legal complexity and user understanding.
Key Features
- Advanced LLM Technology: High-end AI for accurate, relevant legal advice.
- Extensive Legal Knowledge Base: SECP Acts-trained for diverse legal insights.
- User-Friendly Interface: Intuitive Angular-based design for all user levels.
- Robust Backend System: Reliable, scalable performance with Python and Flask.
- Vector Database with RAG: Cutting-edge tech for precise legal query handling.
- Accessibility: Simplifies corporate law for experts and novices alike.
- Conversational and Efficient: Quick, clear legal guidance in a chat format.
Technology Used:
Angular,Flask,Chroma-db,Python,MySQl
Supervisor Name:
Dr.Mehreen Alam
Group Members:
Hammad Anjum (i20 - 0699)
Ammar Haider (i20 - 0513)
Usman Ahmed (i20 - 0820)
LLM-Konnect
LLM-Konnect is an innovative platform designed to empower users to create customized
chatbots leveraging Large Language Models (LLMs). This project enables users to
incorporate their own data sources, integrate pre-built and custom tools, and utilize
automated data analytics for generating insightful graphs and reports. With features
including Text to Speech and Speech to Text capabilities, and the ability to integrate
external services, LLM-Konnect offers versatility for various applications. Users can
choose between a user-friendly dashboard for a no-code chatbot creation experience or
opt for the Software Development Kit (SDK) available on NPM for a more tailored, code-
driven development. Additionally, LLM-Konnect supports IFrame for seamless integration
into existing web infrastructures, making it an all-encompassing solution for individuals
and businesses looking to harness the power of chatbot technology and LLMs for
enhanced communication and data analysis.
Features:
• Customized Chatbots (Knowledge Base + Realtime Data)
• Automated Data Analytics (Graphs + Reports Generation)
• External Services Integration
• Custom Tools Integration
• Text to Speech and Speech to Text
Technology Used:
Langchain, ReactJS, FastAPI, Chroma DB,
OpenAI, Hugging Face, NPM
Supervisor Name:
Dr. Adnan Tariq
Group Members:
Abdul Nafay (i20 - 0492)
Contact #: 0334-6550556
Asadullah Nawaz (i20 - 0761)
Contact #: 0345-0495367
Qasim Ali (i20 - 0766)
Contact #: 0313-7558886
Mentorly
Mentorly is an online mentorship and learning platform built on Microservices Architecture
and is designed to connect mentees with experienced mentors across various domains,
offering personalized guidance and diverse learning experiences. With a focus on user-
centricity, Mentorly facilitates one-on-one mentorship sessions, collaborative problem-
solving forums, industry insights, on-platform evaluation, and webinars. The platform
aspires to become a global hub for effective mentorship and meaningful learning
interactions.
Key Features include:
-Review and Feedback system from mentees
-Public Forums for discussion and error solving
- One on One/Classroom Video Mentoring
- Tier Subscription System
- Virtual Whiteboard/interactive features for sessions
- Mid-session tasks for mentees provided by mentors
-AI Search Engine/ Recommender system
- Payment Management / Community volunteer option
-Webinars
-Industry Insights
Technology Used:
MERN stack, Python, Sentence
Transformers, Flask, GitHub, Flutter,
VSCode
Supervisor Name:
Dr. Khubaib Amjad Alam
Group Members:
Muhammad Usman (i20 - 0416)
Contact: 0310-
8802363
Muhammad Ahmed (i20 - 0498)
Contact:0349-0989199
Muhammad Hamza Shahzad (i20 - 0796)
Contact:0336-9884546
Mind Sight
The project focuses on creating a system for psychologists that enhances their
understanding of people's emotions. It utilizes computer technology to decipher your state
and gain insights into your thoughts. During your visits to the psychologist, you are provided
with devices such as cameras, heart rate monitors and blood pressure trackers. These
gadgets record data while you answer questions and discuss your feelings.
What makes this system unique is its ability to analyze your expressions and listen to your
voice. Assess your bodily signals to gauge your emotional well-being. This helps identify
signs of anxiety or stress which are tremendously valuable. The project facilitates
comprehension for both patients and psychologists regarding what individuals are
experiencing.
With the aid of AI and machine learning algorithms this system becomes highly proficient
in comprehending emotions and their underlying causes. Consequently, psychologists can
make decisions and provide even greater assistance. This project represents an
advancement towards ensuring individuals receive mental health support while
simultaneously simplifying the psychologist's role.
Technology Used:
Tensorflow, MongoDB, Flask, React, Node
js, Python, Jupyter Notebook
Supervisor Name:
Ms. Tajwar Mehmood
Group Members:
Muhammad Ali Raza (i18 - 0570)
Contact: +92 308 5290807
Umer Qazi (i20 - 0968)
Contact: +92 306 0253822
Daniyal Imran (i20 - 0940)
Contact:  +92 303 5053501
MosqueConnect
The MosqueConnect App is a revolutionary app whose aim will be to help you catch your prayer
on times in the Mosque. The users will be able to upload picture of the Mosque clock so that our
app can extract and update the prayer times for that mosque. After that users will verify whether
the time shown is correct or not. Furthermore, it will also broadcast the Azan from each Mosque
from the user’s vicinity and the user will be able to choose from the list of Mosques that they want
to listen to the Azan from. Also, users can use it to listen to the Friday prayer khutba of the Mosques
in their vicinity. They can choose the Mosque and then listen to it on their phones. The app can
also store/download these sermons on cloud or machine and then it will categorize these sermons
according to their topic and give a summary of the khutba as well. The users can add a Mosque
that’s not available on the map as well.
Features include:
• Finding Nearby Mosque
• Getting Prayer Times of Mosque
• Extracting prayer times from the image of a Mosque clock.
• Listening to Azan and Friday Sermons
• Categorizing sermons according to their topics and extracting their summary.
• Adding new Mosques
Technology Used:
MERN, Python, Raspberry PI, Git,
Google Maps API
Supervisor Name:
Mr. Shams Farooq
Group Members:
Mustafa Khalid (i19 - 1964)
Muhammad Hamza (i19 - 0729)
Muhammad Saad Khan (i20 - 2324)
Move: A Posture Correction Fitness App
Move is a fitness app which aims to make everyone’s fitness journey safe and accessible.
Incorrect posture can be the cause of various injuries, hence the app mainly focuses on
analyzing and correcting the user’s exercise form. This is done through professional
trainers, who can use the app to add workouts, and videos of themselves performing
exercises. The users can then pick a trainer of their choosing, and perform those
exercises themselves. The app will compare the user’s form with the trainer’s by
extracting skeletons from the videos, and provide feedback accordingly.
Features include:
- Posture Detection & Correction, using video comparison.
- Extensive Library of trainer uploaded workouts
- Customized Workout Recommendations, based on user-selected goals
- Workout Tracking
- User Progress/Goal Tracking
Technology Used:
Python, MediaPipe, FastAPI, React Native,
Expo, Firebase, Supabase, Figma,
OpenCV
Supervisor Name:
Dr. Adnan Tariq
Group Members:
Muhammad Umar (i20 - 0518)
+92 345 0026695
Abdullah Umar (i20 - 0444)
+92 312 2166999
Aleena Adil (i20 - 2362)
+92 302 2536728
Trust Is Must
Project “Trust is Must” is a Trust-Based rating system with Machine Learning to detect Fake
Reviews which aims to create a comprehensive sentiment analysis platform for product
reviews. It involves web scraping, data preprocessing, and advanced sentiment analysis
techniques to provide users with trustworthy and informed decision-making insights. This
project strives to improve credibility in the realm of online product reviews. It not only builds
trust with consumers but also equips businesses with valuable insights from customer
feedback. By understanding what customers think, companies can improve their products
and create better experiences. This comprehensive understanding helps refine offerings
and boost customer satisfaction. Ultimately, "Trust is Must" aims to create a future where
honest, trustworthy, and data-driven reviews guide confident choices for both consumers
and businesses, ensuring transparency and reliability in the digital marketplace
Features include:
- A user-friendly interface where users can input product URLs to extract relevant review
data.
- The system employs advanced algorithms to deliver precise sentiment scores, allowing
users to accurately distinguish the positivity or negativity of reviews.
- It delves deeper into review insights, offering aspect-based sentiment analysis of
products.
Technology Used:
Python, React, MongoDB, JavaScript,
GitHub
Supervisor Name:
Mr. Jawad Hassan
Group Members:
Talha Zain - (i19 - 0620) – (0340-5555550)
Hasnat Rasool - (i20 - 1833) – (0303-
7364710)
Kamil Ilyas - (i20 - 2371) – (0315-1578275)
NeuraSight - Empowering Banks to Stay Ahead of Fraud
NeuraSight is a modern desktop application designed for financial managers to detect
fraudulent activities effortlessly within transactional data. With advanced features like
Graph Neural Networks, Temporal Motif Analysis, and Time Series Analysis, it swiftly
identifies patterns while providing a comprehensive overview of financial data. Enjoying a
user-friendly interface and intuitive UX design, NeuraSight ensures ease of use, making
fraud detection seamless and easy.
Key features:
1. Graph Neural Networks which dynamically analyze intricate financial networks,
identifying suspicious nodes and connections indicative of money laundering
schemes.
2. Temporal motifs which track recurring patterns over time, uncovering irregularities
in transactional behavior often associated with illicit activities.
3. Time Series Analysis which delves into transactional trends, detecting anomalies
and sudden deviations that may signal money laundering attempts.
Technology Used:
Electron, React, Python, PyTorch,
NetworkX
Supervisor Name:
Dr. Muhammad Arshad Islam
Group Members:
Ahmed Iqbal (i20 - 0447)
03177007705
Mizrab Sheikh (i20 - 0453)
03492113474
Hissam Savul (i20 - 0780)
03329083717
NoCode-ML
An all-in-one, step-by-step, parallelized, and GUI-assisted platform for AI Engineers, Data
Scientists, and Computer Scientists to create and develop machine learning pipelines
easily, without the need to write any code.
Compared to traditional code-based tools, this platform provides a seamless, finite-step
process that allows distributed cloud training of multiple machine learning models in an
automated way with support for quick performance visualization and model inferencing.
The platform also provides a custom GUI-assisted, data-centric pre-processor that can
utilize client-side CPU/GPU parallelism to provide live data feedback. The tool aims to
cater to the growing need to simplify the machine-learning process and provide an easy-
to-use and scalable platform for all sorts of users.
Main Features:
- Easy-to-use Interface: GUI-assisted, user-friendly interface with live data preview.
- Parallelized Pre-processing on client-side: Support for CPU/GPU parallelism on the front
end.
- Distributed Model Training: Enqueue multiple models for distributed training without any
code.
- Seamless Cloud Transition: Integration with cloud services allows for scalable
deployment.
Technology Used:
Ray, AWS, GPU.js (WebCL), D3.js,
FastAPI, MERN, MUI Core, Git, Github
Organization, Firebase, MongoDB Atlas
Supervisor Name:
Dr. Muhmmad Aleem
Co-Supervisor Name:
Mr. Asif Mehmood
Group Members:
Fasih Abdullah (i20 - 0432)
Cell#: +92-315-5585526
Faraz Razi (i20 - 1866)
Cell#: +92-300-4827776
Zeerak Zubair (i20 - 1770)
Cell#: +92-331-4027772
Orange Line Maintenance System
Abstract
Safety is of utmost importance in the transportation industry, and the lack of proper
maintenance protocols directly impacts it. Instances of equipment failures, unexpected
breakdowns, and suboptimal repair practices have led to service disruptions, delays, and,
in some unfortunate cases, severe accidents resulting in injuries or loss of life.
The goal of this project is to streamline the company operations and reduce the workload
and effort needed to carry out daily tasks. This will allow the employees to increase the
productivity of the company, in turn boosting passenger confidence and increasing
ridership.
Features
• Wheel Analysis
• Wheel Data OCR
• Maintenance Record Keeping
• Spare Parts Prediction
• Repetitive and Trending Faults
• Permit to Work
• Comprehensive Dashboard
• Reporting and Analysis
Technology Used:
Flutter, PostgreSQL, Python, Flask, Git
Supervisor Name:
Mr. Saad Salman
Group Members:
Zuha Umar (i20 - 0603)
0347-7651126
Waleed Bin Osama (i20 - 2440)
0336-4208338
Fatima Abdul Wahid (i20 - 0711)
0334-8114321
PathoSync: Enhancing Pathology with Synchronized AI
PathoSync aims to revolutionize digital pathology by introducing an innovative platform
designed to transform how pathologists interact with medical imaging. The project's core
objective is to provide a user-friendly application tailored specifically for pathologists,
empowering them with capabilities for precise image annotation and deep learning without
requiring prior AI expertise. Through intuitive design and advanced functionalities, PathoSync
seeks to enhance diagnostic accuracy, streamline workflows, and democratize AI integration
in pathology.
User-Friendly Interface: Prioritizes ease of use for pathologists, enabling efficient workflow.
Image Pre-processing: Automated techniques enhance image quality for precise annotations.
Image Visualization: Provides interactive visualization options for pathology images.
Cellular Annotations: Allows marking and labeling of individual cells for accurate analysis.
Tissue Annotations: Enables annotation and  of larger tissue regions for improved diagnostics.
Whole Slide Image Annotation: Facilitates annotation and labeling of slide image patches.
Model Training and Finetuning: Enables pathologists to train and finetune custom models for
cell detection, tissue classification, and whole slide label prediction.
Technology Used:
React, Django, MongoDB, Jupyter
Notebook, Tensorflow, Python, AWS,
OpenCV, Keras, OpenSlide
Supervisor Name:
Dr. Mohsin Bilal
Dr. Akhtar Jamil (Co-Supervisor)
Group Members:
Saamiya Mariam (i20 - 0612)    +92 304
5441643
Bisma Haroon (i20 - 0716)    +92 305
5550065
Laiba Imran (i20 - 0991)   +92 303 8700333
Prostate Cancer Detection Using Deep Learning
Prostate cancer is one of the most prevalent cancers among men globally. Early detection
plays a crucial role in effective treatment and improved patient outcomes. Magnetic
Resonance Imaging (MRI) has emerged as a promising modality for prostate cancer
detection due to its high resolution and soft tissue contrast. This project aims to develop a
deep learning-based system for automated prostate cancer detection leveraging MRI
images. The system will incorporate classification and segmentation algorithms to
accurately identify cancerous regions. Additionally, a basic user interface (UI) will be
provided for seamless interaction and visualization of results. Features include:
-MRI Image Processing
-Classification Model
-Segmentation Model
-User Interface (UI)
-Result Visualization
Technology Used:
Tensorflow, PyTorch, Keras, OpenCV,
CUDA, SimpleITK, nilabel, Python, Numpy,
Pandas
Supervisor Name:
Dr. Akhtar Jamil
Group Members:
Saad Bin Farooq (i20-0555)
Mobile: +923092349306
Umer Mukhtar (i20-0696)
Mobile: +923235497349
PDDA: Post-Disaster Damage Assessment
The Web Application would take input images of specific areas and, using segmentation
techniques, would output the damages done to the roads and buildings in a visual format.
This visualized data would also show the degree of damage done to the affected areas by
highlighting and labeling damage accordingly. Each instance of damage assessment, via
the Web Application, will save any related data in the database for accessibility.
Furthermore, the newly generated images can be highlighted and labeled with tools
available in the Web Application so that the status of the damaged areas would be updated
as per user input. Furthermore, the app will be containerized along with its related
dependencies in order to achieve scalability.
There are 3 views in this app:
Admin View: All functions of the application can be accessed with this view. From
adding users to accessing core features of the application. The admin has command
over the whole scope of the application.
Project Lead View: This view is restricted to the project that the relative Project Lead
is in charge of. Each project in the application will have its own Project Lead View.
Team View: This view is for all the members in a specific team. A team is assigned
to map segments within a project and therefore, can only access the core features
related to the assigned segment.
Core features include:
Commenting and highlighting tools for users, project status and recovery updates, report
generation, user/project/team management, visual display of damage done to buildings and
roads.
Technology Used:
MongoDB, Docker, Kubernetes, REACT,
Node.js, NumPy, Tensorflow, Scikit-learn,
Keras
Supervisor Name:
Dr. Akhtar Jamil
Group Members:
Ali Raza (20I-0782)
Ph#:0317 5031906
Hammad Umar (19I-2157)
Ph#:0309 7458770
Polo Champions
Polo Champions is a 3D multiplayer mobile sports game developed in Unity, aiming to
introduce polo to a global audience by delivering competitive gameplay, multiplayer
functionality, and extensive customization options. With engaging gameplay and immersive
graphics, Polo Champions fills a void in the market left by existing polo-themed games.
Players can sign up, create profiles, access features like the shop and leaderboard, and
engage in online matches or play against AI opponents. After matches, players can view
statistics and earn rewards, while also customizing players, horses, and equipment using
in-game currency.
Features include:
• Multiplayer Gameplay
• AI vs Player Mode
• Shop and Customization
• Global Leaderboard
• User Authentication
• Daily Missions and Rewards
• Match Statistics
• Game Settings
• Sound Effects and Music
Technology Used:
C#, Unity, Photon, PlayFab, Blender,
GitHub, Photoshop, Visual Studio
Supervisor Name:
Mr. Bilal Khalid Dar
Group Members:
Shayan Faisal (i20 - 0436)
0335-5025558
Aiza Farooq (i20 - 0688)
0307-0514773
Naik Ur Rehman (i20 - 0871)
0316-7979978
Route Optima
Problem: Inefficient last-mile delivery processes result in delayed deliveries, dissatisfied
customers, and excessive operational costs for courier companies.
Solution: Introducing Route Optima, a cutting-edge software solution designed to
revolutionize the entire last-mile delivery experience for managers, delivery personnel, and
customers alike.
Key Features:
● Smart Route Optimization: Route Optima intelligently plans delivery routes by
considering essential factors like customer time windows, distance, and vehicle
capacity, ensuring maximum efficiency.
● User-Friendly Web Portal: Couriers benefit from a centralized hub that simplifies
route planning, rider assignments, live tracking, delivery status updates, emergency
response, and performance analysis.
● Rider’s App: The Rider App equips delivery personnel with everything they need,
including trip details, navigation assistance, digital receipts, real-time updates,
emergency alerts, and performance insights.
Technology Used:
Node.js, React,Flutter
Python,Docker,Git,Github, Trello
Supervisor Name:
Dr. Arshad Islam
Group Members:
Hammad Habib (i20 - 0864)
Contact: 03157552498
Muhammad Abdullah Cheema (i20 - 0468)
Contact: 03055741696
Manahil Faisal (i15 - 0683)
Contact: 03317770723
SafeGuardHER
SafeGuardHER is a cross-platform mobile application revolutionizing women's safety
during public transportation and travel scenarios. Using cutting-edge audio analysis and
machine learning, the app autonomously detects threats like screams and hate speech in
real-time through ensemble modeling techniques, ensuring rapid and accurate threat
identification. Upon detection, SafeGuardHER triggers instant emergency responses,
including sending SOS messages with precise location data to pre-defined contacts and
initiating calls to nearby police stations. In addition to its advanced threat detection
capabilities, SafeGuardHER boasts innovative features such as a voice-activated SOS
function. Users can trigger emergency alerts simply by uttering the pre-defined keyword
"SafeGuard help”. Another standout is the fake call simulator, strategically designed to
deter potential threats. This functionality generates simulated calls, mimicking official
police communications to create the illusion of assistance and deterrence.
Technology Used:
Flutter, Android Studio, Python, AWS,
Docker, TensorFlow, GitHub
Supervisor Name:
Dr. M Arshad Islam
Group Members:
Saba Karim (i20 - 0813)
Cell: 03365008261
Hamza Iqbal (i20 - 0856)
Cell: 03161439569
Hamza Jadoon (i20 - 0570)
Cell: 03105416397
ScholarMate
Our project delves into the realm of artificial intelligence to efficiently generate concise and
accurate summaries of academic papers. We are rigorously evaluating models such as LED
Large, Pegasus, and BART to identify the most effective one. To ascertain the quality of our
summaries, we employ straightforward metrics such as cosine similarity and ROUGE
scores. Furthermore, we adopt a multi-model strategy, incorporating a zero-shot classifier,
SciBERT, and the optimal summarizer model, to create comprehensive literature reviews
from collections of research papers submitted by users.
Key features:
1. Efficient Summarization: Utilizing advanced AI models like LED Large, Pegasus, and
BART to generate concise and accurate summaries of academic papers.
2. Model Evaluation: Rigorous testing of different models to identify the most effective one
for text summarization. Employing metrics such as cosine similarity and ROUGE scores to
evaluate the quality of the generated summaries.
3. Multi-Model Strategy for Literature Review: Combining a zero-shot classifier,
SciBERT, and the best summarizer model to create comprehensive literature reviews from
user-submitted research papers.
4. Enhanced SciBERT: Training SciBERT on a diverse range of research summaries to
improve its ability to identify crucial details in scholarly articles.
Technology Used:
React, Huggingface, Python, Flask,
MySQL, Docker
Supervisor Name:
Dr. Akhtar Jamil
Group Members:
Zaema Khurram (i20 - 2387) 0335-
9880333
Usman Ali Bokhari (i20 - 0794) 0316-
9217770
Muhammad Raheel (i20 - 0620) 0336-
0558185
Family Vista
Family vista is an innovative web application designed to revolutionize the way families
connect, preserve their legacies, and create lasting memories.
Family vista is a dynamic web application that empowers users to explore and preserve
their family heritage. With interactive family trees, personalized profiles, and time capsule
messages, users can document their family's history in a meaningful and engaging way.
Geographically mapped histories allow users to trace their family's journey across time and
place, while suggested connections help them discover and connect with distant relatives.
The platform also facilitates the sharing of memories through collaborative albums, where
family members can contribute photos, videos, and stories. Real-time notifications keep
users updated on family events and milestones, while cookbooks allow them to preserve
and share cherished family recipes.
Uncover the richness of your family's narrative with Family vista's intuitive features,
fostering deeper connections across generations. Experience the joy of reliving cherished
memories and creating new ones with your loved ones, all within the secure embrace of
Family vista.
Technology Used:
React js, Node js, Express Js, Mongo DB
4.6, Visual Studio Code
Supervisor Name:
Dr. Asif Muhammad Malik
Group Members:
M. Arbaz Ishfaq (i20 - 0521)
+923457380605
Ahmad Nasim (i18 - 0465)
+923334202079
M. Salman Iqbal (i18 - 0607)
+923035190083
Semantic Product Search
An ecommerce website has been made. We’ve assessed the performance of semantic
search algorithms in accurately interpreting user queries and providing relevant product
results and examined the impact of allowing descriptive queries on user interactions and
evaluates the integration of AI-driven technologies in enhancing the shopping journey. Then
based on the system's ability to understand user intent and deliver personalized results.
Semantic Searching – semantically searching the users query against catalog products
and returns results.
• Product Ranking – ranks most to least close match products when it returns
• Word Auto-completion – suggests what the word user is trying to write is
• Spell Correction – fixes any spelling mistakes made by the user
Technology Used:
Python React, TensorFlow, Node js,
MongoDB, Bitbucket, Google Cloud
Platform, Vertex AI, OpenSearch, GitHub,
Visual Studio Code
Supervisor Name:
Dr. Akhtar Jamil
Group Members:
Raheimeen Sherafgan (i20- 0427)
0331-5558122
Fatima Aamir (i20 - 0654)
0303-3881717
Tehreem Arbab (i20 -2481)
0335-5532796
SwiftFix
SwiftFix is an advanced platform designed to revolutionize video editing with its automated
and user-friendly features. It simplifies the editing process by offering tools such as audio
cutting, which optimizes speech by eliminating filler words and pauses. Additionally,
SwiftFix enhances accessibility with automatic subtitle generation, ensuring
comprehension for diverse audiences. The platform also provides prompt thumbnail
design tools, allowing users to effectively capture viewer attention. Another notable feature
is its video summarization capability, condensing lengthy content into concise overviews.
SwiftFix prioritizes audio-visual enhancement, delivering high-quality output with improved
visuals and audio clarity. Its intuitive interface caters to users of all levels, making
professional-grade editing accessible to everyone. With its focus on simplicity and
efficiency, SwiftFix streamlines the editing process, saving users valuable time and effort.
Technology Used:
Figma, VS Code, React.js, Python, Flask,
Tensorflow, PyTorch,MongoDB,
Supervisor Name:
Ms. Hina Binte Haq
Co-Supervisor:
Mr. Saad Salman
Group Members:
Ahmed Mudasar (20i-0497)
0345-5001166
Alina Aftab (20i-0967)
0336-5053497
Humna Arshad (20i-2383)
0304-3928276
Share Ride
Share Ride revolves around the creation of an online user-friendly mobile application that
seamlessly connects riders traveling on a similar route and allows them to share their ride.
Within this application, users are able to register themselves, make real-time ride requests,
assign nearby available drivers, provide the best route and fare estimation, and provide
valuable feedback through review submissions. ShareRide also performs real-time behavioral
monitoring of the driver using digital image processing techniques to ensure safe rides. We
have three main users namely riders, drivers and admin.
Features include:
User Registration - Ride Creation and Booking - Driver Assignment - Ride Sharing - Ride
Management - Ride Tracking and navigation - Fare Calculation - Payment Integration -
Reviews and Ratings - Ride notifications and alerts - Safety and Emergency Features
integration - Driver monitoring - Admin Panel.
Technology Used:
Flutter, Dart, Google Maps,Firebase,
OpenCV, Android Studio
Supervisor Name:
Ms. Saba Kanwal
Group Members:
Rohail Zubair (20I-0463)
Cell #  03335552539
Usman Afzal (20I-0786)
Cell #  03411071580
Abdullah Naseer (20I-0472)
Cell #  03115114010
ShotPerfect – A virtual batting coach
In the fast-paced world of cricket, perfecting your shot can be a challenge, especially when
access to a personal coach is limited. “ShotPerfect” is here to change that narrative, offering
a game-changing solution for batsmen seeking immediate and personalized feedback on
their technique. ShotPerfect works by first allowing cricket players to upload their videos,
focusing on their batting performance. The platform then uses advanced computer vision
and artificial intelligence algorithms to analyze the videos, identifying key elements such as
shot type, pose, and timing. Based on this analysis, ShotPerfect provides tailored feedback
to the players, highlighting areas for improvement and suggesting personalized coaching
tips. This process enables players to refine their batting skills, enhance their technique, and
ultimately improve their performance on the field.
Features include:
• Classifying different types of shots i.e., pull, defense, cover drive etc.
• Analysis of shots indicating potential mistakes.
• Providing personalized feedback on how to improve their batting technique.
• Tracking your improvements and progress
• An intuitive interface with these features integrated.
Technology Used:
Next.js, Flask, Python, Git, Ultralytics,
OpenCV, Google Colab
Supervisor Name:
Dr. Atif Jilani
Group Members:
Sheheryar Ramzan (i20 - 0441)
0314-5428364
Dania Jawad (i20 - 0569)
0333-5202544
Farquleet Farhat Gondal (i20 - 0621)
0305-5106363

AI enrichment