Semantic Segmentation on ACDC Rain (test)
71Learning AccuracyDINO + AutoAlb + CN
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DINO + AutoAlb + CNSSL Pre-training=DINO, Augmentation=AutoAlb, Normalisation=CN2023.03 | 71 | 8 | — | |
| Replay + DINOSSL Pre-training=DINO2023.03 | 70.8 | 2.1 | — | |
| Replay + AutoAlb + CNAugmentation=AutoAlb, Normalisation=CN2023.03 | 69.9 | 3.8 | — | |
| EWC + AutoAlb + CNAugmentation=AutoAlb, Normalisation=CN2023.03 | 69.1 | 4.9 | — | |
| EWC + DINOSSL Pre-training=DINO2023.03 | 66 | 11.3 | — | |
| Replay2023.03 | 64.2 | 2.9 | — | |
| EWC2023.03 | 61.4 | 21.2 | — | |
| ROSETTAtest-time adaptation=true2026.04 | — | — | 58.92 | |
| TENTtest-time adaptation=true2026.04 | — | — | 58.68 | |
| TENT + UniEnttest-time adaptation=true2026.04 | — | — | 58.68 | |
| TENT + UniEnt+test-time adaptation=true2026.04 | — | — | 58.59 |