Table 1 Accuracy of AngioNet using K-fold cross validation.
From: AngioNet: a convolutional neural network for vessel segmentation in X-ray angiography
Dice score | Sensitivity | Specificity | AUC | Pixel accuracy | |
|---|---|---|---|---|---|
k-fold test | 0.856 ± 0.004 | 0.913 ± 0.013 | 0.987 ± 0.001 | 0.991 ± 0.002 | 0.982 ± 0.004 |
k-fold holdout | 0.857 ± 0.012 | 0.909 ± 0.012 | 0.987 ± 0.001 | 0.990 ± 0.002 | 0.980 ± 0.003 |
All data | 0.864 | 0.918 | 0.987 | 0.991 | 0.983 |
