Research Article
A Multiscale-Based Adjustable Convolutional Neural Network for Multiple Organ Segmentation
Table 5
Comparison results of DSC (%) of the proposed method and the state-of-the-art methods on the in-house dataset.
| | L-lung | R-lung | Heart | Sp-Co | Mean |
| U-Net [4] | 96.86 | 96.88 | 94.40 | 89.20 | 94.34 | PSPNet [7] | 96.83 | 96.70 | 94.12 | 83.70 | 92.84 | DeepLab v3+ [11] | 97.25 | 97.01 | 94.10 | 89.07 | 94.36 | DANet [12] | 93.24 | 93.74 | 85.68 | 59.76 | 83.11 | Attention U-Net [5] | 97.24 | 97.10 | 94.36 | 91.09 | 94.95 | Nested U-Net [6] | 96.69 | 96.76 | 94.51 | 89.91 | 94.47 | Ours | 97.76 | 97.52 | 95.27 | 92.24 | 95.70 |
|
|