Automatic Mushroom Species Classification Model for Foodborne Disease Prevention Based on Vision Transformer

Table 2

ViT-L/32 vs. SOTA CNN.

Model	Types	Evaluation metrics (%)
		Precision	Sensitivity	F1-score	ACC	AUC
VGG-16	Macro average	81.85	79.71	80.25	81.31	92.95
	Weighted average	81.59	81.31	81.07
ResNet-34	Macro average	85.64	84.72	84.88	85.79	94.43
	Weighted average	85.85	85.79	85.59
Inception-V3	Macro average	82.21	81.91	81.94	83.04	92.88
	Weighted average	83.01	83.04	82.90
Inception-ResNet-V2	Macro average	85.13	84.29	84.54	85.39	94.42
	Weighted average	85.44	85.39	85.25
Xception	Macro average	92.85	92.61	92.64	92.95	97.82
	Weighted average	93.20	92.95	92.98
ViT-L/32	Macro average	95.69	95.92	95.77	95.97	99.01
	Weighted average	96.05	95.97	95.99