Industrial Anomaly Detection with Skip Autoencoder and Deep Feature Extractor - PubMed Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Nov 30;22(23):9327.
doi: 10.3390/s22239327.

Industrial Anomaly Detection with Skip Autoencoder and Deep Feature Extractor

Ta-Wei Tang  1 , Hakiem Hsu  2 , Wei-Ren Huang  1 , Kuan-Ming Li  1
Affiliations

Industrial Anomaly Detection with Skip Autoencoder and Deep Feature Extractor

Ta-Wei Tang et al. Sensors (Basel). .

Abstract

Over recent years, with the advances in image recognition technology for deep learning, researchers have devoted continued efforts toward importing anomaly detection technology into the production line of automatic optical detection. Although unsupervised learning helps overcome the high costs associated with labeling, the accuracy of anomaly detection still needs to be improved. Accordingly, this paper proposes a novel deep learning model for anomaly detection to overcome this bottleneck. Leveraging a powerful pre-trained feature extractor and the skip connection, the proposed method achieves better feature extraction and image reconstructing capabilities. Results reveal that the areas under the curve (AUC) for the proposed method are higher than those of previous anomaly detection models for 16 out of 17 categories. This indicates that the proposed method can realize the most appropriate adjustments to the needs of production lines in order to maximize economic benefits.

Keywords: anomaly detection; deep learning; feature extraction; industrial defect detection.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Architectures of: (a) AnoGAN, (b) GANomaly, (c) Skip-GANomaly, and (d) DFR.
Figure 2
Figure 2
Structure of proposed method.
Figure 3
Figure 3
MVTec AD, production line smartphone glass-cover, and furniture wood datasets for industrial inspection.
Figure 4
Figure 4
Segmented masks and heat maps of each category created using the proposed method.
Figure 5
Figure 5
Average AUCs of all categories with: (a) different pre-trained feature extractor; (b) different blocks of ResNeXt101 feature extractor.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Xie S., Girshick R., Dollár P., Tu Z., He K. Aggregated Residual Transformations for Deep Neural Networks; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR); Honolulu, HI, USA. 21–26 July 2017; pp. 1492–1500.
    1. Chen Y., Li J., Xiao H., Jin X., Yan S., Feng J. Dual Path Networks. arXiv. 20171707.01629
    1. Tan M., Le Q. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks; Proceedings of the 36th International Conference on Machine Learning; Long Beach, CA, USA. 9–15 June 2019; pp. 6105–6114.
    1. Avdelidis N.P., Tsourdos A., Lafiosca P., Plaster R., Plaster A., Droznika M. Defects Recognition Algorithm Development from Visual UAV Inspections. Sensors. 2022;22:4682. doi: 10.3390/s22134682. - DOI - PMC - PubMed
    1. Huang Y., Xiang Z. RPDNet: Automatic Fabric Defect Detection Based on a Convolutional Neural Network and Repeated Pattern Analysis. Sensors. 2022;22:6226. doi: 10.3390/s22166226. - DOI - PMC - PubMed

LinkOut - more resources