AnomalyVFM -- Transforming Vision Foundation Models into Zero-Shot Anomaly Detectors

Open paper
Benchmark Model Rank Results
zero-shot-anomaly-detection-on-cvc-colondb-zero-shot-cross-domainAnomalyVFM (RADIOv2.5 ViT-L, synthetic-only zero-shot)#2Segmentation AUROC: 85.6Segmentation F1-max: 42.8
zero-shot-anomaly-detection-on-br35h-zero-shot-cross-domainAnomalyVFM (RADIOv2.5 ViT-L, synthetic-only zero-shot)#3Detection AUROC: 94.4
zero-shot-anomaly-detection-on-headct-zero-shot-cross-domainAnomalyVFM (RADIOv2.5 ViT-L, synthetic-only zero-shot)#3Detection AUROC: 94.8Detection F1-max: 90.5
zero-shot-anomaly-detection-on-brainmri-zero-shot-cross-domainAnomalyVFM (RADIOv2.5 ViT-L, synthetic-only zero-shot)#4Detection AUROC: 92.9

Rank counts only results with a code link.