Semantic Segmentation on ACDC online continual (test)
70.9Accuracy (Fog)CoTTA
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CoTTARound=1, Architecture=Segformer-B52022.03 | 70.9 | 41.2 | 62.4 | 59.7 | — | |
| CoTTARound=4, Architecture=Segformer-B52022.03 | 70.9 | 41 | 62.7 | 59.7 | — | |
| CoTTARound=7, Architecture=Segformer-B52022.03 | 70.9 | 41 | 62.8 | 59.7 | — | |
| CoTTARound=10, Architecture=Segformer-B52022.03 | 70.8 | 41 | 62.8 | 59.7 | — | |
| SourceRound=1, Architecture=Segformer-B52022.03 | 69.1 | 40.3 | 59.7 | 57.8 | — | |
| SourceRound=4, Architecture=Segformer-B52022.03 | 69.1 | 40.3 | 59.7 | 57.8 | — | |
| SourceRound=7, Architecture=Segformer-B52022.03 | 69.1 | 40.3 | 59.7 | 57.8 | — | |
| SourceRound=10, Architecture=Segformer-B52022.03 | 69.1 | 40.3 | 59.7 | 57.8 | — | |
| TENT-continualRound=1, Architecture=Segformer-B52022.03 | 69 | 40.2 | 60.1 | 57.3 | — | |
| TENT-continualRound=4, Architecture=Segformer-B52022.03 | 66.5 | 36.3 | 58.7 | 54 | — | |
| TENT-continualRound=7, Architecture=Segformer-B52022.03 | 64.2 | 32.8 | 55.3 | 50.9 | — | |
| BN Stats AdaptRound=1, Architecture=Segformer-B52022.03 | 62.3 | 38 | 54.6 | 53 | — | |
| BN Stats AdaptRound=4, Architecture=Segformer-B52022.03 | 62.3 | 38 | 54.6 | 53 | — | |
| BN Stats AdaptRound=7, Architecture=Segformer-B52022.03 | 62.3 | 38 | 54.6 | 53 | — | |
| BN Stats AdaptRound=10, Architecture=Segformer-B52022.03 | 62.3 | 38 | 54.6 | 53 | — | |
| TENT-continualRound=10, Architecture=Segformer-B52022.03 | 61.8 | 29.8 | 51.9 | 47.8 | — | |
| BN Stats AdaptRound=All, Architecture=Segformer-B52022.03 | — | — | — | — | 52 | |
| CoTTARound=All, Architecture=Segformer-B52022.03 | — | — | — | — | 58.6 | |
| SourceRound=All, Architecture=Segformer-B52022.03 | — | — | — | — | 56.7 | |
| TENT-continualRound=All, Architecture=Segformer-B52022.03 | — | — | — | — | 52.3 |