Project-AgML/taro_leaf_disease_classification

Taro Leaf Disease Classification

Taro (Colocasia esculenta) leaf images labelled healthy, leaf blight, mosaic or root rot, photographed from the dorsal and ventral leaf sides, released with a comparison of Inception V3, ResNet-50 and ViT with SHAP explanations for taro disease identification. The raw config holds the 4,393 original images (all four classes) and the augmented config 21,740 augmented copies of the healthy and leaf-blight images. The leaf side is kept as a column (empty for root rot, which the record does not split by side). The dataset contains raw and augmented versions.
The raw dataset contains 4,393 images.
Images per class:

• Healthy: 464
• Leaf_Blight: 2,255
• Mosaic: 894
• Root_Rot: 780

The augmented dataset contains 21,740 images.
Images per class:

• Healthy: 8,531
• Leaf_Blight: 13,209
• Mosaic: 0
• Root_Rot: 0

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

Citation

@article{maedagutirrez2025ai,
  title={AI-Driven Plant Health Assessment: A Comparative Analysis of Inception V3, ResNet-50 and ViT with SHAP for Accurate Disease Identification in Taro},
  author={Maeda-Gutiérrez, Valeria and Oropeza-Valdez, Juan José and Reveles-Gómez, Luis C. and Padron-Manrique, Cristian and Resendis-Antonio, Osbaldo and Solís-Sánchez, Luis Octavio and Guerrero-Osuna, Hector A. and Olvera Olvera, Carlos Alberto},
  journal={Agronomy},
  volume={15},
  pages={77},
  year={2025},
  publisher={MDPI}
}

The dataset itself can be cited as:

Thite, S., Suryawanshi, Y., & PATIL, K. (2025). Colocasia esculenta Leaf Image Dataset [Dataset]. Mendeley Data. https://doi.org/10.17632/HMDR3DZ3V6.4

This dataset was reformatted from its original format to match HuggingFace standards.