Table 8. Testing results when models run.

1 vs all	Models	Accuracy	Precision	Recall	Auc	F1 score	MCC
MCI/CNAD	EfficientNetB0	0.9705	0.9649	0.9677	0.9680	0.9663	0.9326
	DenseNet121	0.9840	0.9838	0.9796	0.9800	0.9816	0.9633
	AlexNet	0.8827	0.9203	0.8193	0.8190	0.8492	0.7362
AD/CNMCI	EfficientNetB0	0.9880	0.9865	0.9870	0.9870	0.9867	0.9734
	DenseNet121	0.9948	0.9934	0.9951	0.9950	0.9943	0.9885
	AlexNet	0.9958	0.9954	0.9953	0.9950	0.9954	0.9874
CN/MCIAD	EfficientNetB0	0.9757	0.9733	0.9725	0.9730	0.9729	0.9458
	DenseNet121	0.9783	0.9836	0.9685	0.9680	0.9754	0.9665
	AlexNet	0.9654	0.9738	0.9500	0.9500	0.9605	0.9234