Spatially Variant Convolutional Autoencoder Based on Patch Division for Pill Defect Detection
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim, Sora | - |
dc.contributor.author | Jo, Youngjae | - |
dc.contributor.author | Cho, Jungchan | - |
dc.contributor.author | Song, Jiwoo | - |
dc.contributor.author | Lee, Younyoung | - |
dc.contributor.author | Lee, Minsik | - |
dc.date.available | 2020-12-30T01:40:47Z | - |
dc.date.created | 2020-12-21 | - |
dc.date.issued | 2020-12 | - |
dc.identifier.issn | 2169-3536 | - |
dc.identifier.uri | https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/79511 | - |
dc.description.abstract | Detecting pill defection remains challenging, despite recent extensive studies, because of the lack of defective data. In this paper, we propose a pipeline composed of a pill detection module and an autoencoder-based defect detection module to detect defective pills in pill packages. Furthermore, we created a new dataset to test our model. The pill detection module segments pills in an aluminum-plastic package into individual pills. To segment pills, we used a shallow segmentation network that is then divided into individual pills using the watershed algorithm. The defect detection module identifies defects in individual pills. It is trained only on the normal data. Thus, it is expected that the module will be unable to reconstruct defective data correctly. However, in reality, the conventional autoencoder reconstructs defective data better than expected, even if the network is trained only on normal data. Hence, we introduce a patch division method to prevent this problem. The patch division involves dividing the output of the convolutional encoder network into patch-wise features, and then applying path-wise encoder layer. In this process, each latent patch has its independent weight and bias. This can be interpreted as reconstructing the input image using multiple local autoencoders. The patch division makes the network concentrate only on reconstructing local regions, thereby reducing the overall capacity. This prohibits the proposed network reconstructing unseen data well. Experiments show that the proposed patch division technique indeed improves the defect detection performance and outperforms existing deep learning based anomaly detection methods. The ablation study shows the efficacy of patch division and compression following the concatenation of patch-wise features. CCBY | - |
dc.language | 영어 | - |
dc.language.iso | en | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.relation.isPartOf | IEEE Access | - |
dc.title | Spatially Variant Convolutional Autoencoder Based on Patch Division for Pill Defect Detection | - |
dc.type | Article | - |
dc.type.rims | ART | - |
dc.description.journalClass | 1 | - |
dc.identifier.wosid | 000597972600001 | - |
dc.identifier.doi | 10.1109/ACCESS.2020.3041790 | - |
dc.identifier.bibliographicCitation | IEEE Access, v.8, pp.216781 - 216792 | - |
dc.description.isOpenAccess | N | - |
dc.identifier.scopusid | 2-s2.0-85097428524 | - |
dc.citation.endPage | 216792 | - |
dc.citation.startPage | 216781 | - |
dc.citation.title | IEEE Access | - |
dc.citation.volume | 8 | - |
dc.contributor.affiliatedAuthor | Cho, Jungchan | - |
dc.type.docType | Article | - |
dc.subject.keywordAuthor | Anomaly detection | - |
dc.subject.keywordAuthor | Convolutional autoencoder | - |
dc.subject.keywordAuthor | Deep neural network | - |
dc.subject.keywordAuthor | Feature extraction | - |
dc.subject.keywordAuthor | Generative adversarial networks | - |
dc.subject.keywordAuthor | Image reconstruction | - |
dc.subject.keywordAuthor | Licenses | - |
dc.subject.keywordAuthor | Mathematical model | - |
dc.subject.keywordAuthor | Pill defect detection | - |
dc.subject.keywordAuthor | Pipelines | - |
dc.subject.keywordAuthor | Unsupervised learning | - |
dc.subject.keywordPlus | Anomaly detection | - |
dc.subject.keywordPlus | Convolution | - |
dc.subject.keywordPlus | Convolutional neural networks | - |
dc.subject.keywordPlus | Deep learning | - |
dc.subject.keywordPlus | Learning systems | - |
dc.subject.keywordPlus | Pelletizing | - |
dc.subject.keywordPlus | Signal encoding | - |
dc.subject.keywordPlus | Statistical tests | - |
dc.subject.keywordPlus | Anomaly detection methods | - |
dc.subject.keywordPlus | Auto encoders | - |
dc.subject.keywordPlus | Convolutional encoders | - |
dc.subject.keywordPlus | Defect detection | - |
dc.subject.keywordPlus | Detection modules | - |
dc.subject.keywordPlus | Patch divisions | - |
dc.subject.keywordPlus | Plastic packages | - |
dc.subject.keywordPlus | Water-shed algorithm | - |
dc.subject.keywordPlus | Defects | - |
dc.description.journalRegisteredClass | scie | - |
dc.description.journalRegisteredClass | scopus | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
1342, Seongnam-daero, Sujeong-gu, Seongnam-si, Gyeonggi-do, Republic of Korea(13120)031-750-5114
COPYRIGHT 2020 Gachon University All Rights Reserved.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.