Table 1.
Quantitative assessment of convolutional encoder-decoders (U-Net (Ronneberger et al., 2015), v16U-Net, v16pU-Net) embedded with learning schemes P, HP and A over the pathological dataset in Dice, sensitivity, specificity, Jaccard, Cohen’s kappa (%) as well as absolute surface error (mm2). Best results are in bold. Italic underlined scores highlight best results among learning schemes employed with U-Net.
| Metric | Scheme | P | HP | A | ||
|---|---|---|---|---|---|---|
| Network | U-Net (Ronneberger et al., 2015) | v16U-Net | v16pU-Net | |||
| dice ↑ | Deltoid | 68.94 ± 29.9 | 71.05 ± 29.5 | 78.32 ± 24.4 | 80.05 ± 23.1 | 82.42 ± 20.4 |
| Infraspinatus | 71.38 ± 24.7 | 77.00 ± 22.5 | 81.58 ± 18.3 | 81.91 ± 19.0 | 81.98 ± 18.6 | |
| Supraspinatus | 64.94 ± 28.0 | 65.69 ± 29.6 | 65.68 ± 30.7 | 67.30 ± 29.4 | 70.98 ± 28.7 | |
| Subscapularis | 78.10 ± 18.1 | 74.55 ± 25.2 | 81.41 ± 15.0 | 81.58 ± 15.2 | 82.80 ± 14.4 | |
| sens ↑ | Deltoid | 70.85 ± 30.5 | 70.74 ± 29.5 | 78.92 ± 25.4 | 81.45 ± 23.7 | 83.80 ± 21.3 |
| Infraspinatus | 72.12 ± 26.4 | 79.45 ± 23.1 | 84.61 ± 18.2 | 83.74 ± 18.6 | 83.48 ± 19.0 | |
| Supraspinatus | 64.02 ± 31.8 | 63.16 ± 33.2 | 65.55 ± 34.5 | 67.21 ± 33.0 | 68.60 ± 32.3 | |
| Subscapularis | 78.89 ± 19.7 | 74.75 ± 27.3 | 82.53 ± 18.1 | 81.75 ± 18.8 | 84.36 ± 16.5 | |
| spec ↑ | Deltoid | 99.61 ± 0.80 | 99.56 ± 1.07 | 99.85 ± 0.19 | 99.82 ± 0.22 | 99.84 ± 0.22 |
| Infraspinatus | 99.82 ± 0.23 | 99.82 ± 0.22 | 99.84 ± 0.18 | 99.86 ± 0.17 | 99.86 ± 0.18 | |
| Supraspinatus | 99.86 ± 0.18 | 99.90 ± 0.13 | 99.88 ± 0.15 | 99.86 ± 0.17 | 99.91 ± 0.12 | |
| Subscapularis | 99.86 ± 0.13 | 99.83 ± 0.28 | 99.87 ± 0.13 | 99.88 ± 0.12 | 99.86 ± 0.15 | |
| jacc ↑ | Deltoid | 59.27 ± 29.7 | 61.68 ± 29.3 | 69.48 ± 26.0 | 71.46 ± 24.9 | 74.00 ± 22.8 |
| Infraspinatus | 60.32 ± 25.6 | 66.91 ± 24.0 | 72.00 ± 20.4 | 72.63 ± 20.6 | 72.71 ± 21.0 | |
| Supraspinatus | 53.61 ± 27.1 | 55.27 ± 29.3 | 55.70 ± 30.1 | 56.98 ± 28.7 | 61.31 ± 28.7 | |
| Subscapularis | 66.93 ± 19.6 | 64.31 ± 24.7 | 70.83 ± 17.6 | 71.13 ± 17.7 | 72.72 ± 17.16 | |
| kappa ↑ | Deltoid | 68.63 ± 30.0 | 70.73 ± 29.7 | 78.15 ± 24.4 | 79.89 ± 23.2 | 82.28 ± 20.5 |
| Infraspinatus | 71.19 ± 24.7 | 76.85 ± 22.5 | 81.45 ± 18.3 | 81.79 ± 19.0 | 81.86 ± 18.7 | |
| Supraspinatus | 64.76 ± 28.0 | 65.56 ± 29.6 | 65.55 ± 30.7 | 67.16 ± 29.4 | 70.87 ± 28.7 | |
| Subscapularis | 77.95 ± 18.1 | 79.23 ± 17.1 | 81.27 ± 15.0 | 81.45 ± 15.2 | 82.67 ± 14.4 | |
| ASE ↓ | Deltoid | 252.0 ± 421.6 | 268.0 ± 507.8 | 105.5 ± 178.9 | 94.23 ± 139.2 | 80.38 ± 127.5 |
| Infraspinatus | 156.8 ± 228.7 | 92.37 ± 105.9 | 74.47 ± 92.8 | 80.11 ± 96.2 | 79.17 ± 96.9 | |
| Supraspinatus | 174.8 ± 164.0 | 159.9 ± 153.5 | 153.9 ± 146.0 | 147.5 ± 129.4 | 134.6 ± 135.5 | |
| Subscapularis | 94.56 ± 95.5 | 102.0 ± 110.7 | 95.19 ± 109.0 | 94.06 ± 111.3 | 82.95 ± 86.88 | |