Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization
In realizing unsupervised pixel-precise anomaly localization by utilizing a generative model, a reference image must be generated (for comparison with an input image) by transforming abnormal patterns of an input image, if any, into normal patterns. In this study, a patch-level operation with adapti...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2021-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9464236/ |
_version_ | 1811210563836968960 |
---|---|
author | Hyunyong Lee Nac-Woo Kim Jun-Gi Lee Byung-Tak Lee |
author_facet | Hyunyong Lee Nac-Woo Kim Jun-Gi Lee Byung-Tak Lee |
author_sort | Hyunyong Lee |
collection | DOAJ |
description | In realizing unsupervised pixel-precise anomaly localization by utilizing a generative model, a reference image must be generated (for comparison with an input image) by transforming abnormal patterns of an input image, if any, into normal patterns. In this study, a patch-level operation with adaptive patch control is proposed to improve anomaly localization by generating a better reference image. As a way to exploit a generative model, we divide an image into non-overlapped patches of the same size, generate patch-level reference images, and stitch the patch-level reference images into a single reference image. We then conduct anomaly localization by comparing an input image with the stitched, reconstructed image. To effectively apply the patch-level operation, we propose adaptive patch control to determine the number of non-overlapped patches to be applied. For this, we synthesize defective images using normal images and examine how well the candidate methods with different numbers of patches remove the synthesized defects. In the same way, we utilize adaptive patch control to select a promising model among the candidate generative models. Based on experiments conducted using the MVTec Anomaly Detection dataset, we demonstrate that our method outperforms previous existing methods. Under a real-world scenario, our method shows ROC AUC of 0.926, in contrast to the best value of 0.893 from existing studies. Furthermore, we prove the feasibility of the adaptive patch control by showing that the removal of the synthesized defects and the anomaly localization for real defective images are highly correlated. |
first_indexed | 2024-04-12T04:57:59Z |
format | Article |
id | doaj.art-5ddb210a4cf34817ab4bc0d4f3b922ed |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-12T04:57:59Z |
publishDate | 2021-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-5ddb210a4cf34817ab4bc0d4f3b922ed2022-12-22T03:47:04ZengIEEEIEEE Access2169-35362021-01-019907279073710.1109/ACCESS.2021.30919809464236Patch-Level Operation With Adaptive Patch Control for Improving Anomaly LocalizationHyunyong Lee0https://orcid.org/0000-0002-0615-4241Nac-Woo Kim1Jun-Gi Lee2Byung-Tak Lee3Honam Research Center (HRC), Electronics and Telecommunications Research Institute (ETRI), Gwangju, Republic of KoreaHonam Research Center (HRC), Electronics and Telecommunications Research Institute (ETRI), Gwangju, Republic of KoreaHonam Research Center (HRC), Electronics and Telecommunications Research Institute (ETRI), Gwangju, Republic of KoreaHonam Research Center (HRC), Electronics and Telecommunications Research Institute (ETRI), Gwangju, Republic of KoreaIn realizing unsupervised pixel-precise anomaly localization by utilizing a generative model, a reference image must be generated (for comparison with an input image) by transforming abnormal patterns of an input image, if any, into normal patterns. In this study, a patch-level operation with adaptive patch control is proposed to improve anomaly localization by generating a better reference image. As a way to exploit a generative model, we divide an image into non-overlapped patches of the same size, generate patch-level reference images, and stitch the patch-level reference images into a single reference image. We then conduct anomaly localization by comparing an input image with the stitched, reconstructed image. To effectively apply the patch-level operation, we propose adaptive patch control to determine the number of non-overlapped patches to be applied. For this, we synthesize defective images using normal images and examine how well the candidate methods with different numbers of patches remove the synthesized defects. In the same way, we utilize adaptive patch control to select a promising model among the candidate generative models. Based on experiments conducted using the MVTec Anomaly Detection dataset, we demonstrate that our method outperforms previous existing methods. Under a real-world scenario, our method shows ROC AUC of 0.926, in contrast to the best value of 0.893 from existing studies. Furthermore, we prove the feasibility of the adaptive patch control by showing that the removal of the synthesized defects and the anomaly localization for real defective images are highly correlated.https://ieeexplore.ieee.org/document/9464236/Anomaly localizationgenerative modeldeep learningpatch-level operationunsupervised learning |
spellingShingle | Hyunyong Lee Nac-Woo Kim Jun-Gi Lee Byung-Tak Lee Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization IEEE Access Anomaly localization generative model deep learning patch-level operation unsupervised learning |
title | Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization |
title_full | Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization |
title_fullStr | Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization |
title_full_unstemmed | Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization |
title_short | Patch-Level Operation With Adaptive Patch Control for Improving Anomaly Localization |
title_sort | patch level operation with adaptive patch control for improving anomaly localization |
topic | Anomaly localization generative model deep learning patch-level operation unsupervised learning |
url | https://ieeexplore.ieee.org/document/9464236/ |
work_keys_str_mv | AT hyunyonglee patchleveloperationwithadaptivepatchcontrolforimprovinganomalylocalization AT nacwookim patchleveloperationwithadaptivepatchcontrolforimprovinganomalylocalization AT jungilee patchleveloperationwithadaptivepatchcontrolforimprovinganomalylocalization AT byungtaklee patchleveloperationwithadaptivepatchcontrolforimprovinganomalylocalization |