Skip to main content

Table 8 Segmentation results on CT and MR with 30% labels (mean/std)

From: macJNet: weakly-supervised multimodal image deformable registration using joint learning framework and multi-sampling cascaded MIND

Modality

Methods

DSC (%)

Hd95 (mm)

Recall (%)

Precision (%)

RVDabs (%)

VOE (%)

CT

Sub-SegNet

93.39 ± 3.31

19.15 ± 16.66

95.63 ± 6.24

91.50 ± 2.07

7.80 ± 4.25

12.24 ± 5.44

macJNet

95.38 ± 0.82

5.97 ± 3.21

96.05 ± 1.93

94.71 ± 1.24

2.53 ± 2.07

8.81 ± 2.19

MR

Sub-SegNet

93.10 ± 2.39

20.95 ± 18.59

95.23 ± 1.93

91.20 ± 4.25

4.75 ± 6.22

12.81 ± 4.05

macJNet

94.12 ± 1.06

6.59 ± 2.14

94.22 ± 2.13

94.36 ± 2.72

3.92 ± 2.15

12.17 ± 1.86

  1. Bold values indicate better results than other methods