Skip to main content

View full-text article in PMC

. Author manuscript; available in PMC: 2011 May 1.

Published in final edited form as: Neuroimage. 2010 Jan 4;50(4):1519–1535. doi: 10.1016/j.neuroimage.2009.12.092

Table 4.

Numerical results for three tissue-based regression, and only GM-based, by RVR and SVR, with four feature extraction methods.

Mean Square Errors
Features		Regional Clusters	LDA	PCA	ISOMAP
Data and Methods		Regional Clusters	LDA	PCA	ISOMAP
GM, WM and CSF	RVR	10.8338	12.0522	12.5166	18.6454
GM, WM and CSF	SVR	11.3696	12.1237	12.8206	18.8331
GM only	RVR	11.5445	12.2853	12.7153	20.4922
GM only	SVR	12.319	12.6366	13.0917	23.6831
Correlation Coefficients
GM, WM and CSF	RVR	0.75775	0.72625	0.71559	0.53958
GM, WM and CSF	SVR	0.75017	0.7244	0.71466	0.52889
GM only	RVR	0.73868	0.72543	0.71978	0.51375
GM only	SVR	0.72429	0.72542	0.71561	0.28605