Semantic Segmentation on HSICity (test)
50.1mIoUCARL-SSL
Evaluation results
| Method | Links | |
|---|---|---|
| CARL-SSLTraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 50.1 | |
| DOFATraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 49.6 | |
| CARLTraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 48.6 | |
| HyveTraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 48 | |
| Camera-specific modelTraining protocol=Pre-trained on Cityscapes and fine-tuned exclusively on HSICity, Backbone=EVA-022025.04 | 44.6 | |
| HyperFreeTraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 44.6 | |
| Spectral AdapterTraining protocol=Channel-invariant adaptation, concurrently trained on Cityscapes and HSICity, Backbone=EVA-022025.04 | 43.4 |