PlantDoc Dataset Overview

Dataset Name

PlantDoc: A Dataset for Visual Plant Disease Detection

Dataset Purpose

Benchmark for classification models for visual plant disease detection.

Dataset Content

• Contains 2,598 data points.
• Covers 13 plant species.
• Involves up to 17 disease categories.
• Images were scraped from the Internet and manually annotated over ~300 hours.

Dataset Contribution

• Models trained on this dataset can improve plant disease classification accuracy by up to 31%.
• Aims to lower the barrier for applying computer‑vision techniques to plant disease detection.

Related Literature

• Full paper available at Arxiv and ACM.

Dataset Authors

Davinder Singh, Naman Jain, Pranjali Jain, Pratik Kayal, Sudhakar Kumawat, Nipun Batra

Dataset License

Creative Commons Attribution 4.0 International