mukailaalhshituabr-cyber/livestock-disease-detection

Livestock farming is a major source of food and income in Niger, but diseases often kill animals before farmers can access veterinary care. Our AI-powered system helps farmers detect livestock diseases early using images, enabling faster treatment, reducing animal deaths, minimizing financial losses, and improving rural livelihoods. # AI Livestock Disease Detection System An AI-powered mobile app that helps farmers in Niger detect livestock diseases early by photographing a sick animal and getting an instant, AI-generated diagnosis with a confidence score and next-step recommendation. Link to app: livestock-disease-detection… Overview - Problem: Livestock diseases are often caught too late because veterinary services are far away, causing farmers to lose animals and income. - Solution: A farmer photographs their animal in the app. The photo is first checked to confirm it's actually an animal, then a CNN (Convolutional Neural Network) classifies it into a known disease and gives a confidence score and recommendation. A veterinarian can later review and confirm or correct any AI diagnosis. Architecture ``` React Native App (Expo) │ ├──► Supabase (Auth, Postgres database, image storage...) │ └──► FastAPI service (runs ONLY the CNN models, /api/predict) │ ├─ 1. Animal-detection gate (pretrained ImageNet MobileNetV2) └─ 2. Disease classifier (custom-trained MobileNetV2 head) ``` Authentication, user profiles, farms, animals, and the disease/prediction records all live in Supabase. The Python/FastAPI backend does nothing except receive an image and an auth token, verify the token, run the two-stage model pipeline, and return a prediction, it never stores anything permanently (the app writes the result to Supabase itself). Tech Stack | Layer | Technology | Role | |---|---|---| | Mobile app | React Native + Expo | Camera/gallery capture, UI, navigation | | Navigation | React Navigation (stack + bottom tabs) | Screen routing | | Backend-as-a-Service | Supabase (Auth, Postgres, Storage) | Users, data, images | | AI inference service | FastAPI + Uvicorn (Python) | Exposes `/api/predict` | | AI framework | TensorFlow / Keras | Builds and runs the CNNs | | AI model | MobileNetV2 (transfer learning) | Disease classification | | AI safeguard | MobileNetV2 (pretrained …