ログイン 新規登録
言語:

WEKO3

  • トップ
  • ランキング
To
lat lon distance
To

Field does not validate



インデックスリンク

インデックスツリー

メールアドレスを入力してください。

WEKO

One fine body…

WEKO

One fine body…

アイテム

  1. 論文誌(ジャーナル)
  2. Vol.63
  3. No.8

Fixing the Train-test Objective Discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection

https://ipsj.ixsq.nii.ac.jp/records/219095
https://ipsj.ixsq.nii.ac.jp/records/219095
a8eba382-37c7-4595-ac9f-40cc2a0d9216
名前 / ファイル ライセンス アクション
IPSJ-JNL6308004.pdf IPSJ-JNL6308004.pdf (3.3 MB)
Copyright (c) 2022 by the Information Processing Society of Japan
オープンアクセス
Item type Journal(1)
公開日 2022-08-15
タイトル
タイトル Fixing the Train-test Objective Discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection
タイトル
言語 en
タイトル Fixing the Train-test Objective Discrepancy: Iterative Image Inpainting for Unsupervised Anomaly Detection
言語
言語 eng
キーワード
主題Scheme Other
主題 [一般論文] anomaly detection, anomaly segmentation, autoencoder, deep learning, unsupervised learning
資源タイプ
資源タイプ識別子 http://purl.org/coar/resource_type/c_6501
資源タイプ journal article
著者所属
The University of Tokyo
著者所属
The University of Tokyo
著者所属
The University of Tokyo
著者所属(英)
en
The University of Tokyo
著者所属(英)
en
The University of Tokyo
著者所属(英)
en
The University of Tokyo
著者名 Hitoshi, Nakanishi

× Hitoshi, Nakanishi

Hitoshi, Nakanishi

Search repository
Masahiro, Suzuki

× Masahiro, Suzuki

Masahiro, Suzuki

Search repository
Yutaka, Matsuo

× Yutaka, Matsuo

Yutaka, Matsuo

Search repository
著者名(英) Hitoshi, Nakanishi

× Hitoshi, Nakanishi

en Hitoshi, Nakanishi

Search repository
Masahiro, Suzuki

× Masahiro, Suzuki

en Masahiro, Suzuki

Search repository
Yutaka, Matsuo

× Yutaka, Matsuo

en Yutaka, Matsuo

Search repository
論文抄録
内容記述タイプ Other
内容記述 Autoencoders have emerged as a popular method for unsupervised anomaly detection, but they still have difficulty detecting local anomalies in real-world images due to a lack of modeling fine details. We have assessed this difficulty from a new perspective: a mismatch of training and testing objectives. Specifically, we expect autoencoders to encode an unseen locally anomalous image, reconstruct normal regions completely, and repair abnormal parts during testing, even though they merely aim to minimize total reconstruction errors during training. To address this issue, we reconstruct a potentially anomalous masked region from encoding a potentially normal unmasked region conditionally with a mask, similarly to image inpainting, during both training and testing. Because the ideal mask for anomalies is unknown in advance, we iteratively construct an adaptive mask from an earlier anomaly score of the reconstruction error. Our proposed Iterative Image Inpainting for Anomaly Detection (I3AD) updates image inpainting and masking by turns, which engenders the expected objective to maximize the anomaly score during testing. Evaluated by the MVTec Anomaly Detection dataset, our method outperformed baseline reconstruction-based methods in several categories and demonstrated remarkable improvement, especially in high-frequency textures.
------------------------------
This is a preprint of an article intended for publication Journal of
Information Processing(JIP). This preprint should not be cited. This
article should be cited as: Journal of Information Processing Vol.30(2022) (online)
DOI http://dx.doi.org/10.2197/ipsjjip.30.495
------------------------------
論文抄録(英)
内容記述タイプ Other
内容記述 Autoencoders have emerged as a popular method for unsupervised anomaly detection, but they still have difficulty detecting local anomalies in real-world images due to a lack of modeling fine details. We have assessed this difficulty from a new perspective: a mismatch of training and testing objectives. Specifically, we expect autoencoders to encode an unseen locally anomalous image, reconstruct normal regions completely, and repair abnormal parts during testing, even though they merely aim to minimize total reconstruction errors during training. To address this issue, we reconstruct a potentially anomalous masked region from encoding a potentially normal unmasked region conditionally with a mask, similarly to image inpainting, during both training and testing. Because the ideal mask for anomalies is unknown in advance, we iteratively construct an adaptive mask from an earlier anomaly score of the reconstruction error. Our proposed Iterative Image Inpainting for Anomaly Detection (I3AD) updates image inpainting and masking by turns, which engenders the expected objective to maximize the anomaly score during testing. Evaluated by the MVTec Anomaly Detection dataset, our method outperformed baseline reconstruction-based methods in several categories and demonstrated remarkable improvement, especially in high-frequency textures.
------------------------------
This is a preprint of an article intended for publication Journal of
Information Processing(JIP). This preprint should not be cited. This
article should be cited as: Journal of Information Processing Vol.30(2022) (online)
DOI http://dx.doi.org/10.2197/ipsjjip.30.495
------------------------------
書誌レコードID
収録物識別子タイプ NCID
収録物識別子 AN00116647
書誌情報 情報処理学会論文誌

巻 63, 号 8, 発行日 2022-08-15
ISSN
収録物識別子タイプ ISSN
収録物識別子 1882-7764
戻る
0
views
See details
Views

Versions

Ver.1 2025-01-19 14:55:35.283794
Show All versions

Share

Mendeley Twitter Facebook Print Addthis

Cite as

エクスポート

OAI-PMH
  • OAI-PMH JPCOAR
  • OAI-PMH DublinCore
  • OAI-PMH DDI
Other Formats
  • JSON
  • BIBTEX

Confirm


Powered by WEKO3


Powered by WEKO3