Table 1 Comparison of the proposed model with recent state-of-the-art models.

From: Deep intelligence: a four-stage deep network for accurate brain tumor segmentation

 

Loss

Dice coefficient

Jaccard similarity

Tversky index

Accuracy

ARU-GD29

0.4069

0.9189

0.8499

0.7018

0.9979

2D-VNet34

0.0069

0.9192

0.9579

0.9930

0.9937

MultiResUNet35

0.2510

0.7490

0.6004

0.7423

0.9972

3D-UNet36

0.2586

0.7408

0.5900

0.7293

0.9964

2D UNET MODEL39

0.0732

0.9268

0.8803

0.9117

0.9664

LinkNet37

0.0062

0.9938

0.9213

0.9590

0.9942

3D-UNET38

0.4737

1.0

0.250

1

0.25

TransUnet40

0.5891

0.4515

0.2917

0.4515

0.4515

Proposed model (4-stage 2D VNet++)

0.0061

0.9928

0.9964

0.9974

0.9971

Back to article page