Table 3 Comparison of the Proposed Architectures with State-of-the-Art Segmentation Techniques.
From: Lung_PAYNet: a pyramidal attention based deep learning network for lung nodule segmentation
Model | DSC (%) | IOU (%) | Sensitivity (%) | Precision (%) |
|---|---|---|---|---|
Proposed Lung_PAYNet | 95.7 | 91.75 | 92.57 | 96.75 |
UNET | 81.94 | 69.4 | 74.12 | 86.28 |
Central focused CNN (Wang et al.)24 | 82.15 | 71.16 | 92.75 | 75.84 |
3D UNET with LBP, Sobel, Canny operators (Qin et al.)25 | 84.83 | – | 85.11 | 88.95 |
Cascaded dual pathway residual network (Liu et al.)26 | 81.58 | – | 87.3 | 79.71 |
Dual branch residual Network with central intensity pooling (Cao et al.)27 | 82.74 | – | 89.35 | 79.64 |
Dual branch UNET with region growing algorithm (Wu et al.)28 | 83.16 | – | 88.51 | 78.98 |
