Table 2 Quantitative comparison on BraTS2020 dataset.

From: Innovative multi-class segmentation for brain tumor MRI using noise diffusion probability models and enhancing tumor boundary recognition

Method

DICE (%) ↑

HD95 (mm) ↓

WT

TC

ET

Mean

WT

TC

ET

Mean

TransBTS5

90.0

81.9

78.9

83.6

5.64

6.05

3.73

5.14

2D Dense-UNet29

88.1

78.9

71.2

79.4

6.74

10.2

40.6

19.2

nnU-Net6

91.2

85.2

79.5

85.3

3.79

7.77

29.2

13.6

Diff-UNet25

92.2

86.9

76.9

85.4

2.59

3.60

3.98

3.39

Ours

92.4

86.2

77.4

85.3

3.65

5.56

12.9

7.37

Ours (revised ET)

92.4

90.7

88.1

90.4

3.65

4.64

3.09

3.79

  1. Significant values are in bold.
Back to article page