Table 2 Evaluation results of different models in ISLES 2022 segmentation results.

From: SFMANet: A Spatial-Frequency multi-scale attention network for stroke lesion segmentation

Method

DSC

Precision

Recall

F1score

MIoU

UNet

0.5910

0.6182

0.6653

0.5515

0.4892

MDA-Net

0.7044

0.7530

0.7222

0.7368

-

Acc_unet

0.7639

0.8150

0.7379

0.6675

0.6675

UNet++

0.6980

0.7940

0.7028

0.6784

0.5954

U2Net

0.7325

0.7838

0.7062

0.7186

0.6401

AttentionUNet

0.7352

0.8001

0.7110

0.7240

0.6394

SwinUNet

0.5199

0.5949

0.5980

0.5031

0.4150

TransUNet

0.6272

0.7666

0.5842

0.6272

0.5330

TransFuse

0.6872

0.6739

0.7933

0.6868

0.6739

Polyp_PVT

0.7000

0.7054

0.7318

0.6888

0.5802

HmsU-Net

0.6597

0.8037

0.8844

0.6307

0.5679

MSCA-Net

0.7385

0.8633

0.8608

0.7302

0.6343

NLIE-UNet

0.7457

0.8923

0.8923

0.7170

0.6487

Ours

0.7767

0.8784

0.9071

0.7160

0.6911

Back to article page