Table 1 Comparative results of prediction on the testing set (%)
From: Self-relabeling for noise-tolerant retina vessel segmentation through label reliability estimation
Dataset | DRIVE(R) | STARE(VK) | CHASE | ||||
|---|---|---|---|---|---|---|---|
Group | Method | F1 | PR | F1 | PR | F1 | PR |
\(\mathrm {LV}\)-1 | Baseline | 73.2 | 76.0 | 75.7 | 77.7 | 81.9 | 82.8 |
Cas | 76.1 | 83.5 | 72.6 | 78.8 | 79.9 | 87.8 | |
SF | 78.4 | 87.8 | 78.0 | 87.1 | 85.5 | 93.9 | |
Ours | 79.6 | 88.6 | 79.9 | 88.4 | 85.9 | 94.1 | |
\(\mathrm {LV}\)-2 | Baseline | 70.2 | 72.8 | 72.3 | 74.2 | 77.2 | 78.3 |
Cas | 75.1 | 82.0 | 72.9 | 79.8 | 77.4 | 85.4 | |
SF | 75.7 | 84.5 | 77.3 | 86.0 | 83.6 | 92.2 | |
Ours | 77.6 | 87.1 | 78.9 | 87.3 | 84.3 | 92.6 | |
\(\mathrm {LV}\)-3 | Baseline | 67.2 | 69.6 | 69.4 | 71.3 | 73.8 | 74.9 |
Cas | 75.1 | 82.3 | 71.1 | 76.9 | 77.7 | 85.4 | |
SF | 73.2 | 82.4 | 75.0 | 83.5 | 82.4 | 91.0 | |
Ours | 77.0 | 86.4 | 77.6 | 86.4 | 83.1 | 91.4 | |
Pseudo | Baseline | 79.3 | 87.2 | 76.0 | 83.9 | / | / |
Cas | 75.7 | 82.8 | 74.0 | 81.3 | / | / | |
SF | 80.0 | 88.1 | 76.7 | 84.8 | / | / | |
Ours | 80.3 | 88.5 | 76.5 | 84.2 | / | / | |
Manual | Baseline | 78.9 | 79.9 | 72.2 | 76.0 | 76.3 | 78.2 |
Cas | 78.1 | 85.8 | 73.4 | 79.8 | 78.8 | 81.8 | |
SF | 82.8 | 91.2 | 76.7 | 84.1 | 82.7 | 90.4 | |
Ours | 83.0 | 91.3 | 77.8 | 84.3 | 83.3 | 91.1 | |