Cross-Dataset Performance: Model Accuracy on Cocoa Pod Disease Image Datasets

Henry Techie-Menson's research dataset, last updated April 30, 2026, contains performance metrics for a novel deep learning model on multiple cocoa pod disease image datasets. The 5.5 KB XLS file likely includes accuracy, F1, and PPV scores from cross-dataset evaluations. Results are reported for the Cocoa_Pod_Disease_Gh, Cocoa Diseases (YOLOv4), Black and Borer Pod Rot, Cacao Diseases in Davao, and Coffee and Cocoa datasets.

Use Cases

• Benchmarking new cocoa disease classification models based on the reported cross-dataset accuracy scores.
• Studying model generalization in agricultural AI based on the performance across five distinct cocoa disease datasets.
• Evaluating attention mechanisms like LGF-CBAM for plant disease detection based on the described model architecture and results.

Strengths

• Performance metrics are reported with specific percentages, including 98.95% accuracy on the primary dataset.
• Cross-dataset robustness is demonstrated with results on five distinct image collections.
• Dataset is openly licensed under CC-BY-4.0.

Limitations

• The dataset is very small at 5.5 KB, suggesting it contains only summary statistics, not the underlying image data.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Henry Techie-Menson via figshare

Collection Method

Likely compiled from the evaluation results of a deep learning model (LGF-CBAM with ResNetV2-101) on multiple public image datasets.

Freshness

Last updated 2026-04-30 17:52:06; freshness should be verified.