Modelpublic

alanoood494/ishara-dhad-arabic-sign-language

sourceHugging Facemitupdated 2mo agoView on Hugging Face

0likes

Model Card

license: mit language:

—en
—ar libraryname: pytorch pipelinetag: image-classification ---

Ishara Dhad – Arabic Sign Language Recognition

Overview

Ishara Dhad is an AI-powered Arabic Sign Language Recognition model developed as a Software Engineering Graduation Project.

The system recognizes Arabic Sign Language gestures using MediaPipe hand landmarks and a deep learning LSTM model implemented with PyTorch.

Features

—Arabic Sign Language Recognition
—Real-time Prediction
—MediaPipe Landmark Detection
—LSTM Deep Learning Architecture
—PyTorch Implementation
—FastAPI Integration

Technologies

—Python
—PyTorch
—MediaPipe
—OpenCV
—FastAPI

Model Information

Property	Value
Framework	PyTorch
Model	LSTM
Input	MediaPipe Hand Landmarks
Output	Arabic Sign Labels
Supported Classes	20

Supported Signs

—Baby
—Eat
—Father
—Finish
—Good
—Happy
—Hear
—House
—Important
—Love
—Mall
—Me
—Mosque
—Mother
—Normal
—Sad
—Stop
—Thanks
—Thinking
—Worry

Dataset

The model was trained on a custom Arabic Sign Language dataset containing thousands of processed gesture samples.

Project Purpose

Ishara Dhad aims to improve educational accessibility for deaf and hard-of-hearing students by providing an intelligent Arabic Sign Language recognition platform.

Author

Alanoud Emad Aldwaihi

Software Engineering Student

Graduation Project