Image Counts per Class for Rice Leaf Disease Dataset

Ummer Shakeel's dataset provides image counts per class for a rice leaf disease dataset. The underlying dataset contains 1914 balanced image samples and was used to evaluate five deep transfer learning architectures for disease classification. The dataset was last updated on May 7, 2026.

Use Cases

• Benchmarking deep learning models for image classification based on the described evaluation of five architectures.
• Developing interpretable AI systems for agriculture based on the described use of GrabCut segmentation and contour detection for visualization.
• Training lightweight models for mobile or edge deployment based on the reported performance of MobileNetV2.
• Studying transfer learning effectiveness in plant pathology based on the comparative results across different model architectures.

Strengths

• Dataset is balanced with 1914 image samples, as stated in the description.
• Provides specific model performance metrics, such as InceptionV3 achieving 98.43% test accuracy.
• Released under a CC-BY-4.0 license, facilitating open use and sharing.
• Explicitly supports research reproducibility and data transparency through public availability.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 5.5 KB size indicates this file likely contains only summary statistics, not the actual image data.

Provenance

Source

figshare, author Ummer Shakeel.

Collection Method

Images were systematically processed using Python with TensorFlow and a GPU, as described.

Time Range

null

Freshness

Last updated 2026-05-07 17:40:58; freshness should be verified.

Geography

The research context is Pakistan, a major rice producer, but the specific geographic coverage of the images is not stated.