Towards Explainable Visual Anomaly Detection
Authors: Yizhou Wang, Dongliang Guo, Sheng Li, Yun Fu
Abstract: Anomaly detection and localization of visual data, including images and videos, are of great significance in both machine learning academia and applied real-world scenarios. Despite the rapid development of visual anomaly detection techniques in recent years, the interpretations of these black-box models and reasonable explanations of why anomalies can be distinguished out are scarce. This paper provides the first survey concentrated on explainable visual anomaly detection methods. We first introduce the basic background of image-level anomaly detection and video-level anomaly detection, followed by the current explainable approaches for visual anomaly detection. Then, as the main content of this survey, a comprehensive and exhaustive literature review of explainable anomaly detection methods for both images and videos is presented. Finally, we discuss several promising future directions and open problems to explore on the explainability of visual anomaly detection.
Ask questions about this paper to our AI assistant
You can also chat with multiple papers at once here.
Assess the quality of the AI-generated content by voting
Why do we need votes?
Votes are used to determine whether we need to re-run our summarizing tools. If the count reaches -10, our tools can be restarted.
Some bits of the article are not summarized yet, you can re-run the summarizing process by clicking on the Run button below.
Look for similar papers (in beta version)
By clicking on the button above, our algorithm will scan all papers in our database to find the closest based on the contents of the full papers and not just on metadata. Please note that it only works for papers that we have generated summaries for and you can rerun it from time to time to get a more accurate result while our database grows.
Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.