Glass Segmentation on HSO
89.3IoUL+GNet
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| L+GNettraining=all datasets, binary prediction confidences=true2026.03 | 89.3 | 93.4 | 0.04 | 3.98 | |
| L+GNettraining mode=trained and tested separately on respective datasets, with binary prediction confidences=true2026.03 | 88.1 | 92.8 | 0.045 | 4.5 | |
| GlassWizardtraining=all datasets2026.03 | 87.9 | 94.1 | 0.055 | 5.44 | |
| GlassWizardtraining mode=trained and tested separately on respective datasets2026.03 | 86.7 | 92.9 | 0.062 | 6.06 | |
| VBNettraining mode=trained and tested separately on respective datasets2026.03 | 83.1 | 90 | 0.078 | 7.65 | |
| PGSNettraining mode=trained and tested separately on respective datasets2026.03 | 80.1 | — | 0.089 | 9.08 | |
| GSDNettraining mode=trained and tested separately on respective datasets2026.03 | 78.9 | — | 0.103 | 9.79 | |
| GDNettraining mode=trained and tested separately on respective datasets2026.03 | 78.7 | — | 0.097 | 9.32 | |
| Trans2Segtraining mode=trained and tested separately on respective datasets2026.03 | 78 | — | 0.095 | 9.65 | |
| TransLabtraining mode=trained and tested separately on respective datasets2026.03 | 74.3 | — | 0.123 | 12 |