Arabic Sign Language Real-Time Recognition

Overview

This project is a real-time detection system for Arabic Sign Language made by YOLOv10

Table of Contents

1. Introduction
2. Features
3. Demo
4. Installation
5. Results

Introduction

Arabic Sign Language(ASL): A visual-gestural language used by deaf and hard of hearing communities primarily in our Arabian words. Just like spoken languages

• Dataset: Link Text
• Kaggle Notebook: Link Text

Features

• The ASL detection system is able to detect the gestures by predicting the boundary box of where the gesture is and what it describes as a label
• Works on Real-Time inference
• Utilizes any webcam to apply the inference

Demo

Installation (how to use ?)

• 1- create a conda environment with python version 3.10

 conda create --name arabic_sing_language python=3.10

• 2- activate this environment

 conad activate arabic_sign_language

• 2- install pytorch with supporting GPU Setup NVIDIA GPU for PyTorch

 conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

• 3- install some packages

 conda install -c conda-forge ultralytics -y
 conda install -c conda-forge arabic_reshaper -y
 conda install -c conda-forge python-bidi

• 4- use this command to run the project (make sure the camera is open)

 python Arabic_Sign_Language_Recognition.py

Results

• Metrics

• Predictions