moshi-Kassim/model-for-the-early-detection-of-maize-leaf-diseases-in-tanzania

# Maize Leaf Disease Detection Dashboard Streamlit web dashboard for real-time diagnosis of maize leaf diseases using a trained Custom CNN model. ## Project **Developing a Predictive Deep Learning Model for the Early Detection of Maize Leaf Diseases in Tanzania** Detects four classes: - Healthy - Common Rust (*Puccinia sorghi*) - Northern Leaf Blight (*Exserohilum turcicum*) - Cercospora Leaf Spot (*Cercospora zeae-maydis*) ## Requirements - Python 3.10+ - Trained model: `model/corn_disease_cnn.h5` (included) ## Setup ```bash cd "python dashboard" pip install -r requirements.txt ``` ## Run the dashboard There are **two separate apps**: | Dashboard | File | Port | For | |---|---|---|---| | **User Dashboard** | `user_app.py` | 8501 | Farmers — upload, top disease, management only | | **Admin Dashboard** | `app.py` | 8502 | Technical — detailed diagnosis, metrics, about | **User Dashboard (farmers):** ```bash streamlit run user_app.py ``` Or double-click `run_user.bat` / `run.bat` **Admin Dashboard (technical):** ```bash streamlit run app.py --server.port 8502 ``` Or double-click `run_admin.bat` - User app: `localhost` - Admin app: `localhost` Both apps **auto-load** the AI model on startup. Diagnosis runs **automatically** when you upload an image. ## Usage **User Dashboard** 1. Upload a maize leaf image (JPG/PNG) 2. View the detected disease and recommended management **Admin Dashboard** 1. Open **Diagnose** and upload a real maize leaf image for full probabilities and symptoms 2. Open **Performance** for model evaluation metrics 3. Open **About** for project information > For panel presentation, use **live uploads only** (no pre-loaded sample images). Sample files in `assets/sample_images/` are kept for local developer testing only. ## Project structure ``` python dashboard/ ├── user_app.py # User dashboard (farmers) ├── app.py # Admin dashboard (technical) ├── run_user.bat # Start …