AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

Rethinking Unsupervised Outlier Detection via Multiple Thresholding

Zhonghang Liu, Panzhong Lu, Guoyang Xie, Zhichao Lu, Wen-Yan Lin

ECCV 2024 · Label-efficient and robust learning › OOD and anomaly detection · approach: Anomaly detection with contaminated data · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Multi-T inlier/outlier thresholding to DeepSVDDDeepSVDD · STL-10 (ResNet-50)AUC: 0.622 -> 0.925Improved
add Multi-T to OCSVMOCSVM · STL-10 (ResNet-50)AUC: 0.927 -> 0.957Improved
add Multi-T to ABOD on STL-10ABOD · STL-10 (ResNet-50)AUC: 0.665 -> 0.883Improved
add Multi-T to NeuTraL on STL-10NeuTraL · STL-10 (ResNet-50)AUC: 0.854 -> 0.950Improved
add predicted inliers for manifold learning (Multi-T inliers) (ablation)DeepSVDD + Multi-T · STL-10 (ResNet-50)AUC: 0.643 -> 0.950Improved
add predicted outliers for Shell normalization (ablation)DeepSVDD + Multi-T · STL-10 (ResNet-50)AUC: 0.939 -> 0.950Improved
add predicted inliers to LVAD-S (ablation)LVAD-S + Multi-T · STL-10 (ResNet-50)AUC: 0.965 -> 0.971Improved
add outlier normalization to LVAD-S (ablation)LVAD-S + Multi-T · STL-10 (ResNet-50)AUC: 0.951 -> 0.971Improved
use Multi-T instead of classical k-sigma rule (ablation)thresholding · STL-10F0.1: 0.406 -> 0.840Improved
use Multi-T instead of 1-sigma for inlier-recall metric (ablation)thresholding · STL-10F10: 0.806 -> 0.869Improved
use CLIP instead of ResNet-50 features with Multi-TMulti-T · MIT-PlacesAUC: 0.867 -> 0.974Improved
add Multi-T to compress fitting data for OCSVMOCSVM · 10,000 samples runtimetime (s): 154.8s -> 105.4sImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy