Prashant6307/sign-language-recognition

SignSense — Real-Time ASL Sign Language Recognition

A full-stack, real-time American Sign Language (ASL) alphabet recognition system. It captures hand gestures through a webcam, detects landmarks using MediaPipe, classifies them with a trained neural network, and displays live predictions in the browser — complete with a sentence builder and text-to-speech.

GitHub: https://github.com/Prashant6307/Sign-Language-Recognition-System

Features

• —Live webcam capture with real-time hand-skeleton overlay
• —24 ASL static letters recognized (A-Y, excluding J and Z)
• —Sentence builder with backspace, space, clear, and copy controls
• —Text-to-speech to read the built sentence aloud
• —97.4% test accuracy on 24-letter classification

How to use

Click Start Camera, allow camera access, and start signing ASL letters in front of your webcam.