TABLE IV.

Computational Costs of Different Methods in AD Classification and Clinical Score Regression for a New Testing MR Image

Procedure	VBM	ROI	LMF	M3T	M2TF	MSJL	DM2L
Linear Alignment (C++)	5.00 s	5.00 s	5.00 s	5.00 s	5.00 s	5.00 s	5.00 s
Nonlinear Registration (HAMMER [50])	32.00 min	32.00 min	–	32.00 min	32.00 min	32.00 min	–
Landmark Prediction (Matlab)	–	–	10.00 s	–	–	–	10.00 s
Feature Extraction (Matlab)	4.00 s	3.00 s	5.00 s	3.00 s	3.00 s	3.00 s	–
Feature Selection (Matlab)	–	–	–	0.02 s	0.02 s	0.03 s	–
Classification (Matlab)	0.05 s	0.02 s	0.03 s	0.02 s		0.02 s	0.02 s (Tensorflow [48])
Regression (Matlab)	0.16 s	0.08 s	0.12s	0.08 s		0.08 s
Total Time	~ 32.00 min	~ 32.00 min	~ 20.00 s	~ 32.00 min	~ 32.00 min	~ 32.00 min	~ 15 s