- Seif Sherif Assad Ali
- Sama Ahmed ElSayed
- Yusuf Sobhy Sadek Elmeligy
- George Nemr Mellek Poqtor
- Omar Elsayed Elsayed Mousa
This project focuses on developing a Hand Gesture Recognition System that translates Arabic Sign Language (ArSL) into text prompts using deep learning and computer vision. The goal is to create an accessible tool for individuals who rely on sign language for communication, enhancing inclusivity and bridging the gap between the deaf and hearing communities.
Link to Project Presentation: https://www.canva.com/design/DAGnKgSTjkU/oMdTk6_GvWX8X72XdJCY4Q/view?utm_content=DAGnKgSTjkU&utm_campaign=designshare&utm_medium=link2&utm_source=uniquelinks&utlId=h176d3d2eb1
Millions of people use Arabic Sign Language (ArSL) as their primary means of communication, yet there is a significant gap in accessibility tools tailored to the Arabic-speaking deaf community. This project aims to leverage AI and machine learning to provide a practical solution that can recognize hand gestures and convert them into text in real time.
- Dataset: RGB images of Arabic Sign Language letters.
- Preprocessing: Image cleaning, resizing, augmentation, and normalization.
- Model: Using Vision Transformers after trying different models. (Such as Convolutional Neural Networks (CNNs) for hand gesture classification)
- Output: Conversion of gestures into Arabic text.
- Future Expansion: Potential integration with speech synthesis for text-to-speech conversion.
✔ Develop a robust gesture recognition model trained on Arabic sign language datasets.
✔ Ensure high accuracy in gesture-to-text conversion.
✔ Implement a user-friendly interface for real-time interaction.
✔ Promote accessibility and inclusivity for the deaf and hard-of-hearing community.
- This is the numerical features extracted from the original rgb image dataset, it consists of the positions of 21 landmarks relative to the wrist landmark.
- It's position invariant and scale invariant but not rotation invariant because rotation is considered a feature in this data.
- https://www.kaggle.com/datasets/muhammadalbrham/rgb-arabic-alphabets-sign-language-dataset?resource=download&select=RGB+ArSL+dataset
https://drive.google.com/drive/folders/14_aZ9G3FArZuQPstq-lvlq-59tF3Immg?usp=sharing