Feature Selection Based on Machine Learning in MRIs for Hippocampal Segmentation

Table 9

Details of the 23 features resulting by the backward elimination method using Näive Bayes Classifier.

23 features; Backward elimination	Haralick features			Haar-like features	Statistical features
	Orientation	Coordinate	Mask size	Type	Mask size	Entry
Position						Coordinates
Normalized gray level						Value
Correlation	0	Y	7
Correlation	45	X	3
Correlation	45	X	5
Correlation	45	X	7
Correlation	45	X	9
Correlation	45	Y	9
Correlation	45	Y	5
Correlation	90	Y	9
Homogeneity	135	Z	3
Gradient					5
Gradient					7
Contrast	135	Y	3
Gradient					9
Homogeneity	90	X	9
Gradient					3
Skewness					7
Gradient					5
Gradient					3
Template				1
Skewness					5
Gradient					5

The asterisk indicates the entries also present in the list of SFS features. For Haralick features, the orientation in degrees, reference coordinate, and the size of the cubic mask used are reported. In case of Haar-like features, the entry value indicates the template type used (see Figure 2). For statistical/positional kind, the size of the cubic mask used and/or the self-explained value is listed, depending on the specific feature type. In particular for gradients, the column named Entry indicates the segment of the reference diagonal as shown in Figure 3.