Glass Segmentation on GDD
95.1IoUL+GNet
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| L+GNettraining=all datasets, binary prediction confidences=true2026.03 | 95.1 | 97.4 | 2.3 | 2.33 | |
| GlassWizardtraining=all datasets2026.03 | 93.3 | 96.9 | 3.9 | 3.62 | |
| L+GNettraining mode=trained and tested separately on respective datasets, with binary prediction confidences=true2026.03 | 0.948 | 0.972 | 0.025 | 2.5 | |
| C-LPMoE-utraining mode=trained and tested separately on respective datasets2026.03 | 0.923 | — | 0.039 | — | |
| GlassWizardtraining mode=trained and tested separately on respective datasets2026.03 | 0.921 | 0.961 | 0.041 | 3.86 | |
| GlassSemNettraining mode=trained and tested separately on respective datasets2026.03 | 0.908 | 0.95 | 0.045 | 4.34 | |
| VBNettraining mode=trained and tested separately on respective datasets2026.03 | 0.907 | 0.948 | 0.048 | 4.7 | |
| EBLNettraining mode=trained and tested separately on respective datasets2026.03 | 0.882 | 0.935 | 0.056 | 5.38 | |
| PGSNettraining mode=trained and tested separately on respective datasets2026.03 | 0.878 | — | 0.062 | 5.56 | |
| GDNet-Btraining mode=trained and tested separately on respective datasets2026.03 | 0.878 | 0.939 | 0.061 | 5.52 | |
| GDNettraining mode=trained and tested separately on respective datasets2026.03 | 0.876 | 0.937 | 0.063 | 5.62 | |
| GSDNettraining mode=trained and tested separately on respective datasets2026.03 | 0.875 | 0.932 | 0.059 | 5.71 | |
| Trans2Segtraining mode=trained and tested separately on respective datasets2026.03 | 0.844 | — | 0.078 | 7.36 | |
| TransLabtraining mode=trained and tested separately on respective datasets2026.03 | 0.816 | — | 0.097 | 9.7 | |
| SAMtraining mode=trained and tested separately on respective datasets2026.03 | 0.485 | 0.798 | 0.268 | 26.08 |