Comparative Experiment: Mushroom Images for Classification Model Validation

AWPF-ResNet18, a ResNet18-based model with an Adaptive Window Pyramid Fusion module, achieved a 2.5% accuracy improvement in edible mushroom classification tasks. The dataset likely contains images of mushrooms used to validate this model against state-of-the-art alternatives. Author Xinhai Zhao published the supporting data on figshare in April 2026.

Use Cases

• Benchmarking image classification models based on performance metrics like accuracy and macro F1-score
• Training models for edible mushroom identification to prevent accidental poisoning
• Evaluating feature fusion techniques for handling targets of varying sizes within images
• Comparing model architectures for agricultural cultivation and classification tasks

Strengths

• Model performance metrics are quantified, showing a 2.5% accuracy increase and 7.5% macro precision improvement
• Dataset is openly licensed under CC-BY-4.0
• The experimental methodology is described, involving multiple comparative experiments

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• The dataset is very small at 5.5 KB, suggesting limited scope

Provenance

Source

figshare

Collection Method

Likely gathered as part of a comparative experiment for a computer vision model.

Freshness

Last updated 2026-04-16 17:31:46; freshness should be verified