Image Classification on Multi-Domain Set (PACS, OfficeHome, DigitsDG, miniDomainNet) (10 Labels/Class)
81.8PACS AccuracySemAlign
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SemAlignBackbone=ResNet18, Pre-training=ImageNet, Number of labels per class=10, Evaluation Protocol=SSDG2026.01 | 81.8 | 60.3 | 74.2 | 42.2 | 64.6 | |
| StyleMatchBackbone=ResNet18, Pre-training=ImageNet, Number of labels per class=10, Evaluation Protocol=SSDG2026.01 | 79.7 | 59.8 | 66.1 | 40.6 | 61.6 | |
| DGWMBackbone=ResNet18, Pre-training=ImageNet, Number of labels per class=10, Evaluation Protocol=SSDG2026.01 | 78.8 | 59.6 | 65.6 | 41.3 | 61.3 | |
| FBC-SABackbone=ResNet18, Pre-training=ImageNet, Number of labels per class=10, Evaluation Protocol=SSDG2026.01 | 78.1 | 59.3 | 69.4 | 41.3 | 62 | |
| FixMatchBackbone=ResNet18, Pre-training=ImageNet, Number of labels per class=10, Evaluation Protocol=SSDG2026.01 | 77.1 | 58.2 | 64.5 | 39.8 | 59.9 |