Plant Disease Classification Dataset with 51 Classes and 45,000 Images

A multi-crop plant disease classification dataset aggregated from multiple public repositories, comprising 51 disease and healthy classes. It includes approximately 45,000 original images and was used to benchmark vision transformer and custom CNN models. The dataset was published by Md. Sifat Haque Zidan on figshare in April 2026.

Use Cases

• Benchmarking lightweight vision models for plant disease classification based on the 51-class multi-crop dataset.
• Training models for deployment in resource-constrained settings based on the focus on computational efficiency.
• Analyzing model interpretability for agricultural AI based on the mention of Grad-CAM and Grad-CAM++ analyses.
• Studying the impact of controlled image augmentation on generalization in plant disease diagnosis.

Strengths

• Includes approximately 45,000 original images across 51 classes.
• Dataset was used to achieve high validation and test accuracies (up to 99.66%) in the associated study.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The 5.5 KB file size suggests the dataset file likely contains metadata or summary statistics, not the image files themselves.

Provenance

Source

Aggregated from multiple publicly available repositories.

Collection Method

Curated and expanded through controlled augmentation.

Time Range

null

Freshness

Last updated 2026-04-24 17:24:09; freshness should be verified.

Geography

null