American University of Beirut

Skip Navigation LinksHome > Faculty of Arts and Sciences > Department of Computer Science > Projects

Projects

​​​​​​​​​​​Writing Arabic on a Nonlinear Curve​

This project presents a novel approach for writing Arabic text along a nonlinear curve. Traditional methods rely on Latin languages as a basis; however, due to the nature of written Arabic, with its connecting letters, text rendering often struggles with non-linear paths, resulting in distorted connections. Our solution leverages SVG manipulation and B´ezier curves to accurately position Arabic characters along the curve, maintaining smooth connections and aesthetic appeal.


Student: Firaas Antar Professors advising: Mohamed Kobeissi, Amer Abdo Mouawad (research supervisor)

​​



Labeeb

Labeeb: A machine leaning powered approach for the translation of Lebanese sign language via live visual tracking ​

This study presents a method to develop a Lebanese sign language translator. We used existing datasets as well as curated our own in order to represent the alphabet and gestures involved in pronouncing the most common words in the language. Hand detection was used to collect hand landmarks and a unique way of preprocessing data was implemented. Symbol and gesture recognition were built using a simple neural network. Specifically, gesture recognition requires hand tracking to accurately interpret and process the movements and positions of the hand. We propose a methodology that improves model accuracy and results in better classification of symbols and gestures. After testing on the same metrics with other models, we have achieved a more promising result in terms of model accuracy and model training time. ​

Team members: Zein Shehabeddine, Raed Fidawi Advisor: Prof. Rida Assaf ​



Automated Minutes of Meeting Generator

The proposed project aims to develop a web-based software solution designed to streamline the process of capturing and documenting meeting minutes efficiently. The primary functionalities include converting meeting audio into text using advanced speech-to-text technology and generating Meeting Minutes (MoM) based on a predefined template.

Students: Karim El-Akhras, Ahmad El Hage Ali, Adam Ajour Supervisor: Dr. Haidar Safa ​



BeitLegacy

Capstone Project on the Digitalization of Beiteddine Palace.

The digitalization of Beiteddine Palace is a pioneering cultural preservation project designed to safeguard and promote Lebanon's rich heritage. This project involves creating an intricate 3D model of Beiteddine Palace, allowing users to explore its grandeur and historical significance online. Visitors can navigate through the palace and its interior rooms using either a VR headset or a standard mouse interface. Additionally, the platform offers up-to-date information on the palace, and features a virtual museum where users can view and interact with various artifacts. To enhance the educational experience, an integrated chatbot is available to provide detailed insights and answer questions about the palace and its historical context. This innovative approach not only preserves the cultural legacy of Beiteddine Palace but also makes it accessible to a global audience, fostering greater appreciation and understanding of Lebanon's heritage.

Team Members: Alice Karadjian - Leen El Mir - Mohammad Fayad - Serena Kobeissi Supervisor: Haidar Safa ​



BetterReads: A Bookworm's Dream Social Network

BetterReads is a social network that connects reading enthusiasts worldwide, revolutionizing how they engage with literature. With its sleek, intuitive design, BetterReads makes discovering, tracking, and discussing books a breeze. This modern, all-in-one platform is created for today’s book lovers, seamlessly blending reading, tracking, and social interaction.

Positioning itself as an alternative to industry giants like GoodReads, BetterReads offers a more engaging and convenient user experience. Its robust features cater to the diverse needs of contemporary readers, ensuring a vibrant community and dynamic literary exploration.

Powered by Spring Boot for a rock-solid backend, React Native for a dynamic user interface, and MySQL for secure data storage, BetterReads delivers smooth performance, responsive design, and personalized experiences for every user.

Join the BetterReads community and redefine your reading journey.

Team members: Moustapha Ghandour, Lilit Mushegyan, Rawan Darwich, Hassan Hijjawi, Karim Khalil Advisor: Prof. Rida Assaf ​



​​​​​​​​​​​​​Advancing Endoscopic Surgeries through Machine Learning: Automated Detection and Segmentation of Kidney Stones

​"Winner of the Henri Qais Naccache Undergraduate Research Award.

Kidney stones are a common and painful problem that affects many people worldwide. These stones can develop in the kidneys and cause severe pain, often requiring medical intervention. Despite endoscopic surgeries being used as the gold standard for nephrolithiasis, surgical shortcomings still exist and there is room for improvements. For these reasons, automating the detection and segmentation of kidney stones by developing a machine learning-based software is considered a step forward in the ever-evolving field of endourology. Factors such as dust, blurry images, and various types of stones are some of the challenges undermining the effort. In this project, we use a U-Net model to automatically detect and segment kidney stones during endoscopic surgeries. Our model achieved an accuracy of 98.09% and a dice coefficient of 90.47% on the test set. This technology has potential applications in robotic surgery, enhancing the precision and efficiency of surgical procedures by providing real-time, accurate identification of kidney stones.

Team: Khaled Gheith, Nour Obeid, Joudy Al Ashkar. Advisor: Prof. Rida Assaf. ​



​​​​​​​​​​​​​Code Genius

​"Code Genius" aims to deliver fully functional code by using a test case-driven approach. You provide a prompt and test cases; we generate code, compile it in Docker, and validate it against your tests. If it meets the criteria, we confirm its functionality and return the code. If not, we refine and retry until it works perfectly. This ensures the code you receive is reliable and effective. Code Genius, Supervised by Prof Haidar Safa: Mohammad Chaaban, Ali Baraki, Maher Abbas, Mehdi Fawaz, Mohammad Mahdi Bdeir, and Joseph Yazbeck ​

​

Contact Us

For various questions, please try contacting us via social media first!
read more

Privacy Statement

We take data privacy seriously and adhere to all applicable data privacy laws and regulations.
read more

Copyright and Disclaimer

Written permission is needed to copy or disseminate all or part of the materials on the AUB website.
read more

Title IX, Non-Discrimination, and Anti-Discriminatory Harassment

AUB is committed to providing a safe, respectful, and inclusive environment to all members of its community.
read more