Skip to main content
Springer Nature Link
Account

Table 2 Comparison of the segmentation results of TransXAI and SOTA on the BraTS 2019 validation set.

From: Explainable hybrid vision transformers and convolutional network for multimodal glioma segmentation in brain MRI

Method

DSC

HD95

ET

TC

WT

Avg

ET

TC

WT

Avg

U-Net9

  

0.813

   

19.75

 

3D U-Net14

0.709

0.725

0.874

0.769

5.062

8.719

9.43

7.74

KiU-Net18

0.664

0.706

0.861

0.744

9.418

13.04

12.79

11.75

Res U-Net16

0.667

0.706

0.853

0.742

7.270

9.57

8.55

8.46

MS U-Net39

0.713

0.711

0.865

0.763

8.2465

12.65

9.42

10.11

Attention U-Net40

0.7596

0.772

0.888

0.807

5.202

8.26

7.76

7.07

mmFormer41

0.60

0.73

0.829

0.720

    

V-Net42

0.739

0.766

0.887

0.797

6.131

8.71

6.26

7.03

Starke et al.43

0.710

0.710

0.850

0.757

6.57

10.28

8.85

8.57

TransXAI (Ours)

0.745

0.782

0.882

0.803

4.31

7.90

6.36

6.19

  1. Bold and italic represent the best and second-best within each metric column, respectively.
Back to article page