Out-of-Distribution Detection on CIFAR-10 (ID) vs Places365 (OOD)
0.18FPR@95DynProto
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DynProtoBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 0.18 | 99.95 | |
| AdaNDBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 7.42 | 97.92 | |
| DynProtoBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 25.81 | 94.81 | |
| GLMCMBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 27.16 | 91.25 | |
| CSPBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 28.34 | 91.57 | |
| NegRefineBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 29.02 | 90.84 | |
| EnergyBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 39.98 | 91.82 | |
| MCMBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 40.05 | 88.21 | |
| CADRefBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 43.96 | 90.97 | |
| DICEBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 45.3 | 90.19 | |
| ASH-SBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 48.34 | 88.29 | |
| VIMBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 48.87 | 90.2 | |
| NegLabelBackbone=CLIP-B/16, Method Category=Vision Language Model based Methods2026.04 | 50.15 | 85.63 | |
| OptFSBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 51.26 | 89.25 | |
| MSPBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 62.6 | 88.74 | |
| ReActBackbone=ResNet50, Method Category=Vision Model based Methods2026.04 | 68.95 | 83.36 |