Saved in:
Bibliographic Details
Main Authors: Liznerski, Philipp, Varshneya, Saurabh, Calikus, Ece, Wang, Puyu, Bartscher, Alexander, Vollmer, Sebastian Josef, Fellenz, Sophie, Kloft, Marius
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2402.14469
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914246036553728
author Liznerski, Philipp
Varshneya, Saurabh
Calikus, Ece
Wang, Puyu
Bartscher, Alexander
Vollmer, Sebastian Josef
Fellenz, Sophie
Kloft, Marius
author_facet Liznerski, Philipp
Varshneya, Saurabh
Calikus, Ece
Wang, Puyu
Bartscher, Alexander
Vollmer, Sebastian Josef
Fellenz, Sophie
Kloft, Marius
contents Deep learning-based methods have achieved a breakthrough in image anomaly detection, but their complexity introduces a considerable challenge to understanding why an instance is predicted to be anomalous. We introduce a novel explanation method that generates multiple alternative modifications for each anomaly, capturing diverse concepts of anomalousness. Each modification is trained to be perceived as normal by the anomaly detector. The method provides a semantic explanation of the mechanism that triggered the detector, allowing users to explore ``what-if scenarios.'' Qualitative and quantitative analyses across various image datasets demonstrate that applying this method to state-of-the-art detectors provides high-quality semantic explanations.
format Preprint
id arxiv_https___arxiv_org_abs_2402_14469
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reimagining Anomalies: What If Anomalies Were Normal?
Liznerski, Philipp
Varshneya, Saurabh
Calikus, Ece
Wang, Puyu
Bartscher, Alexander
Vollmer, Sebastian Josef
Fellenz, Sophie
Kloft, Marius
Computer Vision and Pattern Recognition
Machine Learning
Deep learning-based methods have achieved a breakthrough in image anomaly detection, but their complexity introduces a considerable challenge to understanding why an instance is predicted to be anomalous. We introduce a novel explanation method that generates multiple alternative modifications for each anomaly, capturing diverse concepts of anomalousness. Each modification is trained to be perceived as normal by the anomaly detector. The method provides a semantic explanation of the mechanism that triggered the detector, allowing users to explore ``what-if scenarios.'' Qualitative and quantitative analyses across various image datasets demonstrate that applying this method to state-of-the-art detectors provides high-quality semantic explanations.
title Reimagining Anomalies: What If Anomalies Were Normal?
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2402.14469