Presentation 2023-12-20
Anomaly detection by deep support data descriptions with pseudo-anomaly data
Shuta Tsuchio, Takuya Kitamura,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper presents deep support vector data description (DSVDD) with pseudo-anomaly data that generated by generative models. Our method is to improve generalization ability. First, we generate mask images from normal images. Next, using normal and mask images, we train generative adversarial network (GAN) or variational autoencoder (VAE), which generate the pseudo-anomaly data. Using the normal and pseudo-anomaly data, we determines the hypersphere by DSVDD. We evaluate the effectiveness of the proposed methods over the conventional methods with the MVTecAD datasets.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) generative adversarial network / deep support vector data description / pseudo-anomaly data / variational autoencoder
Paper # IBISML2023-34
Date of Issue 2023-12-13 (IBISML)

Conference Information
Committee IBISML
Conference Date 2023/12/20(2days)
Place (in Japanese) (See Japanese page)
Place (in English) National Institute of Informatics
Topics (in Japanese) (See Japanese page)
Topics (in English) machine learning, etc.
Chair Masashi Sugiyama(Univ. of Tokyo)
Vice Chair Toshihiro Kamishima(AIST) / Koji Tsuda(Univ. of Tokyo)
Secretary Toshihiro Kamishima(NTT) / Koji Tsuda(Hokkaido Univ.)
Assistant Yoshinobu Kawahara(Osaka Univ.) / Taiji Suzuki(Univ.of Tokyo)

Paper Information
Registration To Technical Committee on Information-Based Induction Sciences and Machine Learning
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Anomaly detection by deep support data descriptions with pseudo-anomaly data
Sub Title (in English)
Keyword(1) generative adversarial network
Keyword(2) deep support vector data description
Keyword(3) pseudo-anomaly data
Keyword(4) variational autoencoder
1st Author's Name Shuta Tsuchio
1st Author's Affiliation National Institute of Technology, Toyama College(NIT, Toyama college)
2nd Author's Name Takuya Kitamura
2nd Author's Affiliation National Institute of Technology, Toyama College(NIT, Toyama college)
Date 2023-12-20
Paper # IBISML2023-34
Volume (vol) vol.123
Number (no) IBISML-311
Page pp.pp.25-30(IBISML),
#Pages 6
Date of Issue 2023-12-13 (IBISML)