Table 7 5-Fold Cross-Validation Accuracy of All Models on Datasets D1 and D2.
From: An attention enhanced CNN ensemble for interpretable and accurate cotton leaf disease classification
Dataset | Model | Fold 1 | Fold 2 | Fold 3 | Fold 4 | Fold 5 | Mean ± SD |
|---|---|---|---|---|---|---|---|
D1 | SqueezeNet | 0.9798 | 0.9798 | 0.9714 | 0.9753 | 0.9701 | 0.9753 ± 0.0041 |
VGG19 | 0.9740 | 0.9837 | 0.9805 | 0.9798 | 0.9772 | 0.9790 ± 0.0033 | |
ResNet101 | 0.9785 | 0.9876 | 0.9798 | 0.9818 | 0.9883 | 0.9832 ± 0.0040 | |
MobileNetV2 | 0.9824 | 0.9876 | 0.9857 | 0.9863 | 0.9850 | 0.9854 ± 0.0017 | |
EfficientNet-B0 | 0.9922 | 0.9839 | 0.9915 | 0.9883 | 0.9928 | 0.9908 ± 0.0018 | |
ShuffleNetV2 | 0.9902 | 0.9824 | 0.9785 | 0.9850 | 0.9772 | 0.9827 ± 0.0047 | |
CottonLeafNet | 0.9974 | 0.9987 | 0.9928 | 0.9954 | 0.9935 | 0.9956 ± 0.0022 | |
D2 | SqueezeNet | 0.9881 | 0.9844 | 0.9863 | 0.9909 | 0.9872 | 0.9874 ± 0.0021 |
VGG19 | 0.9863 | 0.9918 | 0.9872 | 0.9863 | 0.9881 | 0.9879 ± 0.0020 | |
ResNet101 | 0.9918 | 0.9954 | 0.9909 | 0.9927 | 0.9945 | 0.9930 ± 0.0017 | |
MobileNetV2 | 0.9945 | 0.9954 | 0.9936 | 0.9909 | 0.9936 | 0.9940 ± 0.0017 | |
EfficientNet-B0 | 0.9945 | 0.9936 | 0.9954 | 0.9927 | 0.9945 | 0.9945 ± 0.0013 | |
ShuffleNetV2 | 0.9918 | 0.9881 | 0.9899 | 0.9927 | 0.9963 | 0.9918 ± 0.0028 | |
CottonLeafNet | 1.0000 | 0.9973 | 0.9982 | 0.9982 | 0.9991 | 0.9985 ± 0.0009 |
