Table 1.
Performance of the Machine Learning-Based Segmentation Against Observers on the Training and Validation Set
| Overall (N = 8,076) | Training (n = 6,669) | Validation (n = 1,407) | |
|---|---|---|---|
| Lumen area | |||
| Intersection over union | 0.92 ± 0.05 | 0.92 ± 0.05 | 0.91 ± 0.06 |
| Dice similarity coefficient | 0.96 ± 0.03 | 0.96 ± 0.03 | 0.95 ± 0.03 |
| Correlation coefficient | 0.991 (0.991-0.991) | 0.992 (0.991-0.992) | 0.984 (0.983-0.986) |
| Mean difference | −0.08 ± 0.52 | −0.09 ± 0.51 | −0.05 ± 0.56 |
| Vessel area | |||
| Intersection over union | 0.94 ± 0.04 | 0.94 ± 0.04 | 0.93 ± 0.06 |
| Dice similarity coefficient | 0.97 ± 0.02 | 0.97 ± 0.02 | 0.96 ± 0.03 |
| Correlation coefficient | 0.991 (0.991-0.991) | 0.993 (0.993-0.994) | 0.976 (0.974-0.979) |
| Mean difference | −0.11 ± 0.88 | −0.12 ± 0.78 | −0.02 ± 1.26 |
Values are mean ± SD or mean (95% CI).