gyan-io/GSL-backend

GSL Backend - Ghanaian Sign Language Learning Platform # Experimental AI model for Sign Language Interpretation 1. Participated in the 2025 UNICEF Startup Lab Hackathon, which is AI for Inclusive Learning with the team, Creative Inc. 2. Handled backend development with FastAPI, model integration, real-time testing using OpenCV and iterations. 3. Our team achieved reliable hand gesture recognition with the Sign-to-Text feature and 77.78% accuracy on single words and short phrases. A FastAPI-based backend system that powers an AI-driven mobile application for learning and translating Ghanaian Sign Language. The system enables real-time bidirectional translation between sign language, speech, and text, transforming the static Harmonized GSL dictionary into an interactive learning platform. ## Features - **AI-Powered Sign Recognition**: Computer vision models for real-time GSL gesture recognition - **Speech-to-Sign Translation**: Convert Ghanaian English speech to GSL signs - **Text-to-Sign Translation**: Transform text input into GSL sign demonstrations - **Interactive Learning**: Structured lessons, tutorials, and gamified progress tracking - **Offline-First Design**: Optimized for low-bandwidth environments in rural Ghana - **Cultural Localization**: Support for Ghanaian English accents and local phrases ## Technology Stack - **Backend Framework**: FastAPI 0.104.1 - **Database**: PostgreSQL with SQLAlchemy ORM - **Caching**: Redis for performance optimization - **AI/ML**: TensorFlow Lite, Whisper, OpenCV - **Authentication**: JWT tokens with bcrypt hashing - **Containerization**: Docker and Docker Compose ## Quick Start ### Prerequisites - Python 3.11+ - Docker and Docker Compose - Git ### Installation 1. **Clone the repository** ```bash git clone cd gsl-backend ``` 2. **Set up environment variables** ```bash cp .env.example .env # Edit .env with your configuration ``` 3. **Start with Docker Compose (Recommended)** ```bash docker-compose up -d ``` 4. **Or run locally** ```bash # Install dependencies pip ins …