Anomaly Detection on Fishyscapes Lost & Found
43.9APDenseHybrid
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DenseHybridAux data=true, Img rsyn.=false2022.07 | 43.9 | 6.2 | — | |
| SynBoostAux data=true, Img rsyn.=true2022.07 | 43.2 | 15.8 | — | |
| SynBoostscore=dissimilarity score, re-training=false, OoD data=true, time [s]=2.322019.04 | 43.2 | 15.8 | — | |
| Dirichlet priorAux data=true, Img rsyn.=false2022.07 | 34.3 | 47.4 | — | |
| Dirichlet DeepLabscore=prior entropy, re-training=true, OoD data=true, time [s]=0.052019.04 | 34.3 | 47.4 | — | |
| Standardized MLAux data=false, Img rsyn.=false2022.07 | 31.1 | 21.5 | — | |
| OOD HeadAux data=true, Img rsyn.=false2022.07 | 30.9 | 22.2 | — | |
| Outlier Headscore=combined probability, re-training=true, OoD data=true, time [s]=0.122019.04 | 30.9 | 22.2 | — | |
| Proposed single-reference anomaly detection methodBackbone=DINOv3 [16], Training Protocol=training-free, Reference Setup=single-reference, Resolution=fixed spatial resolution2026.05 | 26.43 | 92.76 | 61.95 | |
| Outlier Headscore=random patches, re-training=true, OoD data=true, time [s]=0.122019.04 | 21.2 | 36.9 | — | |
| Outlier Headscore=fixed patches, re-training=true, OoD data=true, time [s]=0.122019.04 | 15.7 | 76.9 | — | |
| Void ClassifierAux data=true, Img rsyn.=false2022.07 | 10.3 | 22.1 | — | |
| OoD trainingscore=void classifier, re-training=true, OoD data=true, time [s]=0.052019.04 | 10.3 | 22.1 | — | |
| Mutual informationAux data=true, Img rsyn.=false2022.07 | 9.8 | 38.5 | — | |
| Bayesian DeepLabscore=mutual information, re-training=true, OoD data=false, time [s]=3.622019.04 | 9.8 | 38.5 | — | |
| Image Resyn.Aux data=true, Img rsyn.=false2022.07 | 5.7 | 48.1 | — | |
| Image Resynthesisscore=resynthesis difference, re-training=false, OoD data=false, time [s]=2.392019.04 | 5.7 | 48.1 | — | |
| Embed. Dens.Aux data=false, Img rsyn.=false2022.07 | 4.7 | 24.4 | — | |
| Learned Embedding Densityscore=logistic regression, re-training=false, OoD data=true, time [s]=1.532019.04 | 4.7 | 24.4 | — | |
| Learned Embedding Densityscore=minimum NLL, re-training=false, OoD data=false, time [s]=1.532019.04 | 4.3 | 47.2 | — | |
| kNN Embeddingscore=density, re-training=false, OoD data=false, time [s]=7.892019.04 | 3.5 | 30 | — | |
| Learned Embedding Densityscore=single-layer NLL, re-training=false, OoD data=false, time [s]=0.292019.04 | 3 | 32.9 | — | |
| Softmaxscore=entropy, re-training=false, OoD data=false, time [s]=0.292019.04 | 2.9 | 44.8 | — | |
| Softmaxscore=max-probability, re-training=false, OoD data=false, time [s]=0.052019.04 | 1.8 | 44.8 | — | |
| Max softmaxAux data=false, Img rsyn.=false2022.07 | 1.77 | 44.9 | — | |
| OoD trainingscore=max-entropy, re-training=true, OoD data=true, time [s]=0.292019.04 | 1.7 | 30.6 | — | |
| kNN Embeddingscore=relative class density, re-training=false, OoD data=false, time [s]=6.452019.04 | 0.8 | 100 | — | |
| Randomscore=random uncertainty, re-training=false, OoD data=false, time [s]=-2019.04 | 0.3 | 95 | — |