From: DCAU-Net: dense convolutional attention U-Net for segmentation of intracranial aneurysm images
Component | Dice | Sensitivity | Specificity |
---|---|---|---|
  Ours | 0.7455 | 0.7881 | 0.9996 |
  UNet | 0.4629 | 0.6077 | 0.9992 |
  UNet + Dense | 0.5424 | 0.6068 | 0.9995 |
  UNet + CAM | 0.6783 | 0.6254 | 0.9998 |
  UNet + SAM | 0.6426 | 0.7850 | 0.9997 |
  UNet + CBAM | 0.6813 | 0.7398 | 0.9997 |
  UNet + Dense + CBAM | 0.7292 | 0.7378 | 0.9996 |
  Use BN | 0.6612 | 0.6829 | 0.9996 |
  Use ReLU activation | 0.5731 | 0.6560 | 0.9993 |