Table 2.
Summary of the best performance of different coronary vessels segmentation methods.
| Study | Method | Dataset size | Dice | Sensitivity | Specificity | Precision | Accuracy |
|---|---|---|---|---|---|---|---|
| Felfelian et al. (17) | Thresholding | 50 (Test) | 72.79 | 74.92 | 98.32 | – | 97.09 |
| Tsai et al. (24) | Tracking | 20 (Test) | – | 96.70 | 96.30 | – | 96.30 |
| Mabrouk et al. (39) | Graph-cut | 91 (Test) | 75.60 | 76.60 | – | 77.60 | – |
| Lv et al. (46) | Deformable model | 4 (Test) | 76.24 | 72.33 | – | 80.59 | – |
| Jin et al. (52) | PCA | 223 (Test) | 76.97 | 71.25 | 83.95 | – | – |
| Zhu et al. (58) | CNN | 73 (Train), 36 (Test) | 88.40 | 87.30 | – | 90.10 | – |
| Iyer et al. (59) | Encoder-decoder | 370 (Train), 92 (Test) | 86.40 | 91.80 | 98.70 | – | 98.30 |
| Yang et al. (64) | U-Net | 2,642 (Train), 660 (Test) | 89.60 | 89.30 | – | 90.60 | – |
| Hamdi et al. (67) | GAN | 100 (Train), 50 (Test) | 81.18 | 81.09 | 98.11 | 81.26 | 96.55 |
| Tao et al. (70) | Attention mechanism | 104 (Train), 30 (Test) | – | 87.70 | 97.89 | – | 97.29 |
| Gao et al. (74) | Ensemble method | 104 (Train), 26 (Test) | 87.40 | 90.20 | 99.20 | 85.70 | – |