At-an/wastedetect

This project deals with the design and implementation of a low cost object detection system for waste sorting and recycling in Cameroon. Buea as the case study. # WasteDetect A low-cost, human-assisted PWA for real-time waste sorting and recycling built for resource-constrained environments in Buea, Cameroon. ## Tech Stack - **Frontend:** React.js, WebRTC, Chart.js (hosted on Vercel) - **Backend:** Flask, YOLOv8n (hosted on Railway) - **Database:** SQLite (development), PostgreSQL (production) - **Storage:** Cloudinary - **Auth:** JWT (Flask-JWT-Extended) ## Project Structure - `/backend` — Flask API and YOLOv8n model - `/frontend` — React PWA - `/docs` — Diagrams, ERD, and UI designs ## Model Training Pipeline. The model was trained on a 43,869 waste image dataset consisting of five main waste categories locally found in Buea, Cameroon. **For More Details** In case you need more details about the model training and deployment process (data preparation, model training, model evaluation, and model deployment), check it out at kaggle.com