ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables
In many anomaly detection tasks, where anomalous data rarely appear and are difficult to collect, training using only normal data is important. Although it is possible to manually create anomalous data using prior knowledge, they may be subject to user bias. In this paper, we propose an Anomalous La...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9761923/ |
| _version_ | 1818215273613754368 |
|---|---|
| author | Hironori Murase Kenji Fukumizu |
| author_facet | Hironori Murase Kenji Fukumizu |
| author_sort | Hironori Murase |
| collection | DOAJ |
| description | In many anomaly detection tasks, where anomalous data rarely appear and are difficult to collect, training using only normal data is important. Although it is possible to manually create anomalous data using prior knowledge, they may be subject to user bias. In this paper, we propose an Anomalous Latent variable Generative Adversarial Network (ALGAN) in which the GAN generator produces pseudo-anomalous data as well as fake-normal data, whereas the discriminator is trained to distinguish between normal and pseudo-anomalous data. This differs from the standard GAN discriminator, which specializes in classifying two similar classes. The training dataset contains only normal data; the latent variables are introduced in anomalous states and are input into the generator to produce diverse pseudo-anomalous data. We compared the performance of ALGAN with other existing methods on the MVTec-AD, Magnetic Tile Defects, and COIL-100 datasets. The experimental results showed that ALGAN exhibited an AUROC comparable to those of state-of-the-art methods while achieving a much faster prediction time. |
| first_indexed | 2024-12-12T06:33:28Z |
| format | Article |
| id | doaj.art-cf09ff6037874590bcb7662e440ab85a |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-12T06:33:28Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-cf09ff6037874590bcb7662e440ab85a2022-12-22T00:34:32ZengIEEEIEEE Access2169-35362022-01-0110442594427010.1109/ACCESS.2022.31695949761923ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent VariablesHironori Murase0https://orcid.org/0000-0003-1861-6826Kenji Fukumizu1https://orcid.org/0000-0002-3488-2625Department of Statistical Science, The Graduate University for Advanced Studies, SOKENDAI, Tachikawa, JapanThe Institute of Statistical Mathematics, Tachikawa, JapanIn many anomaly detection tasks, where anomalous data rarely appear and are difficult to collect, training using only normal data is important. Although it is possible to manually create anomalous data using prior knowledge, they may be subject to user bias. In this paper, we propose an Anomalous Latent variable Generative Adversarial Network (ALGAN) in which the GAN generator produces pseudo-anomalous data as well as fake-normal data, whereas the discriminator is trained to distinguish between normal and pseudo-anomalous data. This differs from the standard GAN discriminator, which specializes in classifying two similar classes. The training dataset contains only normal data; the latent variables are introduced in anomalous states and are input into the generator to produce diverse pseudo-anomalous data. We compared the performance of ALGAN with other existing methods on the MVTec-AD, Magnetic Tile Defects, and COIL-100 datasets. The experimental results showed that ALGAN exhibited an AUROC comparable to those of state-of-the-art methods while achieving a much faster prediction time.https://ieeexplore.ieee.org/document/9761923/Anomaly detectioncomputer visionmachine learningdeep learninggenerative adversarial networks |
| spellingShingle | Hironori Murase Kenji Fukumizu ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables IEEE Access Anomaly detection computer vision machine learning deep learning generative adversarial networks |
| title | ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables |
| title_full | ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables |
| title_fullStr | ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables |
| title_full_unstemmed | ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables |
| title_short | ALGAN: Anomaly Detection by Generating Pseudo Anomalous Data via Latent Variables |
| title_sort | algan anomaly detection by generating pseudo anomalous data via latent variables |
| topic | Anomaly detection computer vision machine learning deep learning generative adversarial networks |
| url | https://ieeexplore.ieee.org/document/9761923/ |
| work_keys_str_mv | AT hironorimurase algananomalydetectionbygeneratingpseudoanomalousdatavialatentvariables AT kenjifukumizu algananomalydetectionbygeneratingpseudoanomalousdatavialatentvariables |