Table 1.

Average evaluation metrics of 3D-based convolutional networks compared to the traditional 2D-based convolutional networks.

Labels	Classifier	Recall	Precision	F1-score
Contrast	2D	0.61	0.62	0.61
	3D	0.96	0.94	0.95
NCI	2D	0.59	0.54	0.56
	3D	0.93	0.9	0.91
MCI	2D	0.57	0.55	0.56
	3D	0.94	0.93	0.93
Dementia	2D	0.62	0.64	0.63
	3D	0.95	0.91	0.93