Detect
YOLOv8 tracks signing in the live camera feed.
AI/MLBefore Sept 2025
GesturePro: Interactive Sign Language Translator
AI-powered ASL translator bridging signers and non-signers. Uses YOLOv8 object detection with LSTM/TCN for real-time ASL-to-text and speech. Shipped as a PWA built with Next.js, FastAPI and PostgreSQL.
Gallery
Sequence
LSTM/TCN models turn gestures into language.
Speak
Text and speech output for non-signers.
Product
GesturePro translates American Sign Language in real time—detection with YOLOv8, sequence modeling with LSTM/TCN, and a PWA front end so signers and non-signers can communicate without specialized hardware.
Problem
ASL communication barriers persist when interpreters or specialized devices aren’t available.
Solution
A browser-friendly PWA: camera in, text and speech out, backed by FastAPI and PostgreSQL.
Highlights
- Real-time ASL-to-text/speech with YOLOv8 + LSTM/TCN
- Progressive Web App
Specs
- Detection
- YOLOv8
- Sequence models
- LSTM / TCN
- Frontend
- Next.js PWA
- API
- FastAPI
- Data
- PostgreSQL
- Latency goal
- Real time
How it works
- 1Open the PWA and allow camera access.
- 2Sign in frame; YOLOv8 tracks gestures live.
- 3Read the translated text as LSTM/TCN decode the sequence.
- 4Enable speech output when speaking to non-signers.
Achievements
- End-to-end ASL → text/speech stack
- Shipped as an installable Progressive Web App
- Full-stack: Next.js, FastAPI, PostgreSQL
Stack
YOLOv8LSTM/TCNNext.jsFastAPIPostgreSQL
Core StackFrontendBackend & DatabasesCloud & DevOpsData & Machine LearningDesign & DX