IGD (pre-trained ImageNet)

Reported on 5 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology5 results

Anomaly DetectiononMNIST
ROC AUC· uses extra data· 2021-01-25
99.27
best: 99.7 (GAN-based Anomaly Detection in Imbalance Problems)
SOTA
Deep One-Class Classification via Interpolated Gaussian Descriptor arXiv:2101.10043
Anomaly DetectiononFashion-MNIST
ROC AUC· 2021-01-25
93.57
best: 98.6 (GAN based Anomaly Detection in Imbalance Problems)
Deep One-Class Classification via Interpolated Gaussian Descriptor arXiv:2101.10043
Anomaly DetectiononOne-class CIFAR-10
AUROC· 2021-01-25
83.68
best: 99.6 (CLIP (OE))
Deep One-Class Classification via Interpolated Gaussian Descriptor arXiv:2101.10043
Anomaly DetectiononMVTec AD
Detection AUROC· 2021-01-25
92.6
best: 99.9 (GLASS)
Deep One-Class Classification via Interpolated Gaussian Descriptor arXiv:2101.10043
Anomaly DetectiononMVTec AD
Segmentation AUROC· 2021-01-25
91
best: 99.7 (WeakREST-Block)
Deep One-Class Classification via Interpolated Gaussian Descriptor arXiv:2101.10043