Sign Language Recognition AI
This project is a real-time Sign Language (ASL) recognition system using Mediapipe for hand landmark detection and TensorFlow for machine learning.
The model has been converted to TensorFlow.js so it can run directly in a web browser with live webcam input.
Features
- Collect hand sign data for ASL letters
- Train a neural network model to recognize signs
- Live prediction from webcam
- Can be hosted online using GitHub Pages with TensorFlow.js
- Uses lightweight landmark-based AI for fast, real-time performance