Visual Localization on i12Scenes 5° 5 cm threshold (test)
1.1Median Translation Error (cm)SARS
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Kitchen-22026.04 | 1.1 | 0.5 | 97.6 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Bed2026.04 | 1.4 | 0.6 | 93.6 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Kitchen-12026.04 | 1.5 | 0.8 | 92.7 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Living-12026.04 | 1.5 | 0.6 | 97.2 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Living-22026.04 | 1.5 | 0.6 | 98.9 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Floor.5b2026.04 | 1.6 | 0.5 | 99.8 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Gate3622026.04 | 1.7 | 0.8 | 97.9 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Lounge2026.04 | 2.2 | 0.8 | 93.9 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Luke2026.04 | 2.3 | 1 | 86.5 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Manolis2026.04 | 2.5 | 1 | 85.8 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Floor.5a2026.04 | 2.7 | 1.1 | 85.7 | — | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Gate3812026.04 | 3.2 | 1.5 | 71.9 | — | — | |
| Buff-CSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 83.1 | 15.7 | |
| Fine-tuneEvaluation Protocol=Fine-tuning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 6.2 | — | |
| GDREvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 78.7 | 21.9 | |
| GECEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 13.5 | 88.9 | |
| iCaRLEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 66 | 33.3 | |
| JointEvaluation Protocol=Joint Training, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 96.9 | — | |
| SARSEvaluation Protocol=Continual Learning, Backbone=GLACE [13], Threshold=5° 5 cm, Scene=Aggregate2026.04 | — | — | — | 91.8 | 8.1 |