Set Features for Anomaly Detection
Niv Cohen, Issar Tzachor, Yedid Hoshen
TL;DR
This paper introduces set features that model samples by the distribution of their elements, enabling detection of anomalies based on unusual element combinations, outperforming existing methods in image and time series anomaly detection.
Contribution
The paper proposes a novel set feature approach that captures element distributions within samples, addressing limitations of existing methods in detecting anomalies from unusual element combinations.
Findings
Outperforms state-of-the-art in image-level logical anomaly detection.
Outperforms state-of-the-art in sequence-level time series anomaly detection.
Uses simple density estimation with fixed features for anomaly scoring.
Abstract
This paper proposes to use set features for detecting anomalies in samples that consist of unusual combinations of normal elements. Many leading methods discover anomalies by detecting an unusual part of a sample. For example, state-of-the-art segmentation-based approaches, first classify each element of the sample (e.g., image patch) as normal or anomalous and then classify the entire sample as anomalous if it contains anomalous elements. However, such approaches do not extend well to scenarios where the anomalies are expressed by an unusual combination of normal elements. In this paper, we overcome this limitation by proposing set features that model each sample by the distribution of its elements. We compute the anomaly score of each sample using a simple density estimation method, using fixed features. Our approach outperforms the previous state-of-the-art in image-level logical…
Peer Reviews
Decision·ICLR 2024 Conference Withdrawn Submission
1. The usage of set features as presented in the paper (i.e., random projections of feature histograms and Mahalanobis distance scoring) is novel for the anomaly detection task to my best understanding. 2. The method has not only been image but time-series datasets
1. The method is not clearly presented. The notations are unclear; What is f_i[j] in Sec. 3.3? How is f in Eq. (1) exactly obtained? 2. One of the papers main claim that it is the sota on MVTec-Loco, is not valid; the proposed method is outperformed by EfficientAD [2], which is 2 years already outdated. Particularly, the proposed method significantly underperforms in the structural anomaly detection. 3. The method has not been tested on conventional benchmarks MVTec-AD [3] and VisA [4]. 4. There
- The intuition behind the histogram descriptors is good
1. The paper should present the algorithm better. Currently, some vital information is only in the Introduction (last para on page 1). That should be moved to Section 3 for clarity. 2. Section 3.1 second paragraph "The typical way ...": While the discussion here refers to only the pooling aspect of deep networks, it overlooks their property of detecting higher-level abstractions. There are multiple feature maps in the layers of deep nets and each feature map detects one type of abstraction. A
(1) This work tackles the intricate issue of logical anomalies in anomaly detection, a problem not commonly addressed. (2) The paper is clear and easy to follow. (3) The inclusion of experiments covering both image anomaly detection and time-series datasets adds to the paper's breadth and applicability.
(1) The method relies on a strong assumption that individual elements within a sample are normal, but their combination leads to anomalies (i.e., logical anomalies). However, there is no foolproof way to validate this assumption when a query sample is introduced, casting doubt on the practicality of the algorithm. (2) The evaluation lacks comprehensiveness as it does not include state-of-the-art (SOTA) anomaly detection algorithms like CutPaste, RD4AD, SimpleNet, PaDim, CS-Flow, etc., in its ex
- The paper is well written and easy to follow. - Anomaly detection is an interesting and timely topic.
- The method is applicable to a narrow case, specifically when the anomalies are represented by an unusual distribution of normal elements. This setting does not apply to detecting anomalies in general. The proposed method should be integrated within a method able to detect other kinds of anomalies. - There are several methods reporting much better results on MVTec-LOCO (see [1]). The authors claim to achieve state-of-the-art results, but according to [1], this is clearly not the case. Since the
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAnomaly Detection Techniques and Applications · Artificial Immune Systems Applications · Network Security and Intrusion Detection
MethodsSparse Evolutionary Training