Peach Leaf Damage Classification on Peach Leaf Damage Public Datasets (cross-validation)
93.6F1 (W)EfficientNetB5 + CBAM
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| EfficientNetB5 + CBAMImg Size=224, Model parameter scale category=CBAM / Attention-Enhanced Models, Attention-Enhanced (CBAM)=true2026.06 | 93.6 | 93.3 | 83.8 | 87.1 | 84.9 | |
| EfficientNetB0Img Size=224, Model parameter scale category=Light Models (< 15M params), Attention-Enhanced (CBAM)=false2026.06 | 92.7 | 92.9 | 86.1 | 78.6 | 81.7 | |
| DenseNet121Img Size=192, Model parameter scale category=Light Models (< 15M params), Attention-Enhanced (CBAM)=false2026.06 | 92.3 | 92.6 | 84 | 78.7 | 80.1 | |
| EfficientNetB5Img Size=192, Model parameter scale category=Large Models (> 60M params), Attention-Enhanced (CBAM)=false2026.06 | 92.2 | 91.9 | 87.6 | 83.2 | 83.7 | |
| EfficientNetB3Img Size=192, Model parameter scale category=Medium Models (15M – 60M params), Attention-Enhanced (CBAM)=false2026.06 | 92.1 | 91.5 | 82.8 | 81.9 | 80.6 | |
| ResNet101Img Size=192, Model parameter scale category=Large Models (> 60M params), Attention-Enhanced (CBAM)=false2026.06 | 91.9 | 91.9 | 79.5 | 77.9 | 77.8 | |
| EfficientNetB3 + CBAMImg Size=192, Model parameter scale category=CBAM / Attention-Enhanced Models, Attention-Enhanced (CBAM)=true2026.06 | 91.5 | 91.5 | 82.8 | 81.4 | 81.9 | |
| ResNet50Img Size=192, Model parameter scale category=Medium Models (15M – 60M params), Attention-Enhanced (CBAM)=false2026.06 | 91.3 | 90.8 | 82.2 | 78.1 | 76.4 | |
| InceptionV3 + CBAMImg Size=224, Model parameter scale category=CBAM / Attention-Enhanced Models, Attention-Enhanced (CBAM)=true2026.06 | 90.9 | 91.2 | 82.3 | 77.6 | 79.1 | |
| VGG16Img Size=224, Model parameter scale category=Medium Models (15M – 60M params), Attention-Enhanced (CBAM)=false2026.06 | 90.5 | 90.8 | 79.6 | 72.2 | 75 | |
| VGG19Img Size=192, Model parameter scale category=Medium Models (15M – 60M params), Attention-Enhanced (CBAM)=false2026.06 | 90.1 | 89.8 | 81.7 | 78.6 | 78.9 | |
| MobileNetV2Img Size=192, Model parameter scale category=Light Models (< 15M params), Attention-Enhanced (CBAM)=false2026.06 | 89.9 | 89.8 | 77.8 | 73 | 73.3 | |
| MobileNetV3LargeImg Size=224, Model parameter scale category=Light Models (< 15M params), Attention-Enhanced (CBAM)=false2026.06 | 89.9 | 90.1 | 78.9 | 74.3 | 76.4 | |
| InceptionV3Img Size=224, Model parameter scale category=Medium Models (15M – 60M params), Attention-Enhanced (CBAM)=false2026.06 | 88.9 | 89.4 | 86.9 | 70.1 | 76.2 |