🌿 FoodSafe — AI-Powered Food Adulteration Detection

A research-grade, full-stack web application for detecting food adulteration using AI, computer vision, and NLP in Hindi/Marathi/English.

🎯 Project Overview

FoodSafe helps Indian families detect food adulteration using:

• Real-time camera detection (YOLOv8)
• Native Hindi/Marathi NLP (IndicBERT/MuRIL)
• Predictive risk scoring (Prophet time-series)
• Personalized health profiles (scikit-learn)
• FSSAI violation data integration

📁 Project Structure

foodsafe/
├── frontend/          # React.js web app
├── backend/           # FastAPI Python backend
├── ml/                # ML models, notebooks, training scripts
└── docs/              # Research paper, API docs

🛠️ Tech Stack

Layer	Technology
Frontend	React.js, Tailwind CSS, Leaflet.js
Backend	FastAPI, PostgreSQL, Redis, Celery
AI/ML	Claude API, YOLOv8, IndicBERT, Prophet
Hosting	Vercel (FE), Render (BE), Supabase (DB)
Cost	₹0 (all free tiers)

🚀 Quick Start

Prerequisites

• Node.js 18+
• Python 3.10+
• PostgreSQL
• Redis

Frontend

cd frontend
npm install
npm run dev

Backend

cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload

🔬 Research Questions

1. Can multimodal AI detect food adulteration more accurately than single-modality approaches?
2. How does regional NLP (Hindi/Marathi) improve food safety awareness vs English-only?
3. Can time-series ML on FSSAI data predict seasonal adulteration spikes?
4. What is the impact of personalized toxin exposure scoring on dietary behaviour??

📄 License

MIT License — open for research and educational use.