Table 2.
Result 1: performance of the source domain.
| Method | FFHQ | ProGAN | |
|---|---|---|---|
| Ours | Accuracy | 0.994 | 0.989 |
| Precision | 0.990 | ||
| Recall | 0.994 | ||
| F1-Score | 0.992 | ||
| CNN+Self-attention (2020) | Accuracy | 0.991 | 0.974 |
| Precision | 0.984 | ||
| Recall | 0.991 | ||
| F1-Score | 0.987 | ||
| Pupil regular recognition+Boundary IoU score (2022) | Accuracy | 0.995 | 0.988 |
| Precision | 0.988 | ||
| Recall | 0.995 | ||
| F1-Score | 0.992 | ||
| Dual-color spaces+improved Xception model (2021) | Accuracy | 0.994 | 0.981 |
| Precision | 0.981 | ||
| Recall | 0.994 | ||
| F1-Score | 0.988 | ||
| MaskCNN+RAN (2022) | Accuracy | 0.989 | 0.975 |
| Precision | 0.975 | ||
| Recall | 0.989 | ||
| F1-Score | 0.982 | ||