Table 1 Dice and Matthews coefficients for each sample, obtained from the comparison of our neural network results and data from Larson et al.7.
From: A reusable neural network pipeline for unidirectional fiber segmentation
Sample | Tiramisu | U-Net | 3D Tiramisu | 3D U-Net | ||||
|---|---|---|---|---|---|---|---|---|
Dice | Matthews | Dice | Matthews | Dice | Matthews | Dice | Matthews | |
232p1, wet | 97.58 ± 2.29% | 96.55 ± 2.93% | 97.58 ± 2.20% | 96.60 ± 2.13% | 94.54 ± 6.73% | 92.28 ± 9.65% | 95.59 ± 0.74% | 93.71 ± 1.03% |
232p3, cured | 98.21 ± 0.04% | 97.47 ± 0.06% | 98.26 ± 0.04% | 97.53 ± 0.06% | 95.25 ± 6.36% | 93.39 ± 8.88% | 95.90 ± 1.00% | 94.21 ± 1.30% |
232p3, wet | 97.79 ± 2.15% | 96.87 ± 2.70% | 97.85 ± 2.12% | 96.98 ± 1.99% | 94.86 ± 6.90% | 92.76 ± 9.87% | 95.68 ± 1.97% | 93.92 ± 2.36% |
244p1, cured | 98.42 ± 0.03% | 97.83 ± 0.05% | 98.38 ± 0.04% | 97.78 ± 0.05% | 94.55 ± 7.74% | 92.67 ± 10.54% | 96.30 ± 1.25% | 94.97 ± 1.54% |
244p1, wet | 98.08 ± 2.53% | 97.39 ± 3.15% | 98.10 ± 2.39% | 97.43 ± 2.23% | 94.81 ± 7.81% | 92.97 ± 10.71% | 96.67 ± 1.00% | 95.45 ± 1.31% |
