Bilateral Image Subtraction and Multivariate Models for the Automated Triaging of Screening Mammograms

Table 2

Features of the proposed models.

#	-normalization						-rank normalization
	Healthy vs Calcifications			Healthy vs Masses			Healthy vs Calcifications			Healthy vs Masses
	View	Image	Feature	View	Image	Feature	View	Image	Feature	View	Image	Feature
1	CC		27	CC		40	CC		2	CC		40
2	CC		13	CC		29	CC		9	CC		12
3	CC		29	CC		40	CC		12	CC		29
4	CC		29	CC		6	MLO		42	CC		32
5	CC		6	MLO		11	MLO		13	CC		28
6	MLO		28	CC		29	MLO		32	MLO		32
7	MLO		11	CC		28	MLO		6	MLO		13
8	MLO		21	CC		38	CC		2	CC		28
9	CC		28	CC		12	CC		29	CC		12
10	CC		38	MLO		40	CC		14	MLO		2
11	MLO		27	MLO		28	CC		11	MLO		5
12	CC		27	MLO		29	MLO		15	MLO		39
13				MLO		31	MLO		16
14				MLO		7	CC		4
15				MLO		39	CC		39
16				MLO		27	CC		19
17							CC		12
18							CC		17
19							CC		12
20							MLO		31
21							MLO		12

Features are grouped by dataset, symmetric features are denoted with: , , , , , , , , , .