Table 9 Performance Evaluation and Reliability Metrics for Dataset D1 and D2.
From: An attention enhanced CNN ensemble for interpretable and accurate cotton leaf disease classification
Dataset | Model | Cohen’s | Brier | Accuracy 95% CI | Mean | Mean |
|---|---|---|---|---|---|---|
Kappa | Score | [Lower, Upper] | PPV | NPV | ||
D1 | ShuffleNetV2 | 0.9550 | 0.0097 | [0.9375, 0.9833] | 0.9636 | 0.9926 |
SqueezeNet | 0.9600 | 0.0079 | [0.9417, 0.9875] | 0.9668 | 0.9933 | |
VGG19 | 0.9450 | 0.0103 | [0.9250, 0.9792] | 0.9548 | 0.9909 | |
ResNet101 | 0.9500 | 0.0099 | [0.9333, 0.9833] | 0.9608 | 0.9917 | |
MobileNetV2 | 0.9550 | 0.0098 | [0.9375, 0.9833] | 0.9643 | 0.9925 | |
EfficientNet-B0 | 0.9700 | 0.0082 | [0.9542, 0.9918] | 0.9760 | 0.9950 | |
CottonLeafNet | 0.9800 | 0.0051 | [0.9667, 0.9958] | 0.9838 | 0.9967 | |
D2 | ShuffleNetV2 | 0.9848 | 0.0059 | [0.9593, 0.9969] | 0.9889 | 0.9962 |
SqueezeNet | 0.9771 | 0.0073 | [0.9600, 1.0000] | 0.9838 | 0.9944 | |
VGG19 | 0.9619 | 0.0104 | [0.9429, 0.9943] | 0.9724 | 0.9907 | |
ResNet101 | 0.9771 | 0.0075 | [0.9600, 1.0000] | 0.9833 | 0.9944 | |
MobileNetV2 | 0.9848 | 0.0046 | [0.9714, 1.0000] | 0.9885 | 0.9962 | |
EfficientNet-B0 | 0.9848 | 0.0041 | [0.9714, 1.0000] | 0.9885 | 0.9962 | |
CottonLeafNet | 0.9924 | 0.0044 | [0.9829, 1.0000] | 0.9943 | 0.9981 |
