Table 7.

Classification evaluation results (%) of each classifier against the proposed Ensemble CNN + ViT.

Methods	F1-score						F1-score (avg) (%)	Accuracy (avg) (%)
	Cr	In	Pa	Ps	Rs	Sc
Se_resnet50	92.3	96.3	97.5	96.5	95.9	97.6	96	96.3
Resnext50_32 × 4d	94.6	96.6	97.9	94.2	96.1	98.8	96.4	96.7
EfficientnetB3	95.4	96.3	96.5	93.6	99	98.2	96.5	96.7
EfficientnetB1	95.4	96.0	97.9	95.2	96.4	98.2	96.5	96.7
EfficientnetB4	96.4	96.9	97.5	93.8	98.5	96.5	96.6	96.8
EfficientnetB2	96.3	96.9	97.9	95.8	97	97.1	96.8	97.0
EfficientnetB0	97.2	97.0	99.3	95.8	98	95.2	97.1	97.3
Resnet152	98.1	96.9	99.3	97.2	96.5	97.6	97.6	97.7
Wide_resnet	98.1	97.6	99.3	94.5	97.4	98.8	97.6	97.8
Resnet34	94.8	98.3	98.6	97.9	98.4	98.2	97.7	98.0
Densenet121	98.2	97.3	98.6	97.9	98.0	99.4	98.2	98.1
Resnet50	98.2	97.9	98.9	97.2	98.5	100	98.4	98.5
Resnext101_32 × 8d	98.1	98	99.2	97.9	98.1	100	98.5	98.5
EfficientnetB5	98.1	98	98.9	97.9	98.6	100	98.6	98.6
Resnet101	98.1	97.6	98.9	98.6	100	98.2	98.6	98.6
The Proposed Ensemble CNN + ViT	100	99.3	100	100	99	100	99.7	99.6