ai-agriculture-circuits-and-systems/cassava

This project focuses on developing a machine learning model to classify diseases in Cassava plants using leaf images. Cassava is a crucial food security crop in Africa, primarily grown by small-holder farmers. The goal is to help farmers quickly identify plant diseases to prevent crop loss and ensure food security. # Cassava Leaf Disease Classification Cassava leaf disease classification dataset. Cassava is a crucial food security crop in Africa, primarily grown by small-holder farmers. The goal is to help farmers quickly identify plant diseases to prevent crop loss and ensure food security. - **Project page**: `kaggle.com` - **Dataset repository**: `github.com` ## TL;DR - **Task**: Classification (five classes: healthy, cbb, cgm, cmd, cbsd) - **Modality**: RGB - **Platform**: Ground (handheld/field) - **Real/Synthetic**: Real - **Images**: 5,925 labeled images (plus 3,774 unlabeled test images) - **Classes**: 5 classes (healthy, cbb, cgm, cmd, cbsd) - **Resolution**: Various - **Annotations**: COCO JSON (image-level via full-image boxes) - **License**: MIT (see License) - **Citation**: see below ## Table of Contents - Download - Dataset Structure - Sample Images - Annotation Schema - Stats and Splits - Quick Start - Evaluation and Baselines - Datasheet (Data Card) - Known Issues and Caveats - License - Citation - Changelog - Contact ## Download **Original dataset**: `kaggle.com` This repo hosts structure and conversion scripts only; place the downloaded folders under this directory. **Local license file**: See `LICENSE` in the root directory (MIT License). **Alternative sources**: Data hosted on Kaggle. ## Dataset Structure This dataset follows the standardized dataset structure specification with subcategory organization: ``` cassava/ ├── healthy/ # Healthy cassava leaves │ ├── csv/ # CSV annotations per image │ ├── json/ # (empty, no JSON annotations) │ ├── images/ # Healthy images │ └── segmentations/ # (empty, no segmentation masks) ├── cbb/ # Cassava Bacterial Blight │ ├── csv/ │ ├── json/ │ ├── images/ │ └── segme …