Skip to main content

Table 4 Numbers of detected targets for visual and CNN4 analyses

From: Convolutional neural network -based phantom image scoring for mammography quality control

CNN

Visual

f1

f2

f3

f4

f5

f6

c1

c2

c3

c4

c5

m1

m2

m3

m4

m5

nonv

f1

158

0

1

0

0

0

0

0

0

0

0

0

0

0

0

0

2

f2

0

146

0

0

0

0

0

0

0

0

0

0

0

0

0

0

4

f3

0

0

140

0

0

0

0

0

0

0

0

0

0

0

0

0

2

f4

0

0

0

119

0

0

0

0

0

0

0

0

0

0

0

0

2

f5

0

0

0

0

90

0

0

0

0

0

0

0

0

0

0

0

9

f6

0

0

0

1

0

20

0

0

0

0

0

0

0

0

0

0

6

c1

0

0

0

0

0

0

160

0

0

0

0

0

0

0

0

0

0

c2

0

0

0

0

0

0

0

157

0

0

0

0

0

0

0

0

0

c3

0

0

0

0

0

0

0

1

139

0

0

0

0

0

0

0

6

c4

0

0

0

0

0

0

0

0

9

115

0

0

0

0

0

0

3

c5

0

0

0

0

0

0

0

0

0

0

17

0

0

0

0

0

5

m1

0

0

0

0

0

0

0

0

0

0

0

160

0

0

0

0

0

m2

0

0

0

0

0

0

0

0

0

0

0

0

154

1

0

0

4

m3

0

0

0

0

0

0

0

0

0

0

0

0

2

148

4

0

4

m4

0

0

0

0

0

0

0

0

0

0

0

0

0

5

137

1

6

m5

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

85

17

nonv

0

2

1

4

6

5

0

1

0

2

2

0

0

0

0

4

493

  1. f stands for fibre (e.g. “f1” means the first, i.e. the most visible, fibre), c stands for microcalcification group, m stands for mass, and nonv stands for a non-visible target. The CNN4 was allowed to detect multiple possible targets for each input image patch (therefore, e.g., the sum of the first row is greater than 160)