Image Category Discovery on ImageNet 30-class subsets (test)
74.4AccuracyCCN-CCL
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CCN-CCLBackbone=ResNet-18, Pre-training=ImageNet882 classification2018.06 | 74.4 | 71.5 | 0.762 | 0.765 | |
| CCN-KCLBackbone=ResNet-18, Pre-training=ImageNet882 classification2018.06 | 73.8 | 65.2 | 0.75 | 0.715 | |
| LSCBackbone=ResNet-18, Pre-training=ImageNet882 classification2018.06 | 73.3 | 33.5 | 0.733 | 0.655 | |
| K-meansBackbone=ResNet-18, Pre-training=ImageNet882 classification2018.06 | 71.9 | 34.5 | 0.713 | 0.671 | |
| LPNMFBackbone=ResNet-18, Pre-training=ImageNet882 classification2018.06 | 43 | 21.8 | 0.526 | 0.5 |