Color Constancy on NUS-8 (test)
1.83Mean ErrorVLM-CC
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VLM-CCProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 1.83 | 1.44 | 1.5 | 0.51 | 3.88 | |
| CCMNetProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 2.32 | 1.71 | 1.83 | 0.53 | 5.18 | |
| C5 (m = 9)Protocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU, m=92026.05 | 2.54 | 1.9 | 2.02 | 0.61 | 5.61 | |
| C5 (m = 7)Protocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU, m=72026.05 | 2.68 | 2 | 2.14 | 0.66 | 5.9 | |
| FFCCProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 2.87 | 2.14 | 2.3 | 0.71 | 6.23 | |
| Grayness IndexProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 2.91 | 1.97 | 2.13 | 0.56 | 6.67 | |
| PCA-based B/W ColorsProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 2.93 | 2.33 | 2.42 | 0.78 | 6.13 | |
| Quasi-UProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 3 | 2.25 | — | — | — | |
| Cross-dataset CCProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 3.08 | 2.24 | — | — | — | |
| Local Surface Refl.Protocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 3.45 | 2.51 | 2.7 | 0.98 | 7.32 | |
| Shades-of-GrayProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 3.67 | 2.94 | 3.03 | 0.98 | 7.75 | |
| Gray-worldProtocol=Leave-one-out cross-dataset, Train Sets=Gehler-Shi, Cube+, Intel-TAU2026.05 | 4.59 | 3.46 | 3.81 | 1.16 | 9.85 |