The outlier-exposed style distillation network (OE-SDN) mimics the style translation and suppresses the content translation of the autoencoder (AE). Given an input image,...
Abstract:
Autoencoders (AEs) have been widely used for unsupervised anomaly detection. They learn from normal samples such that they produce high reconstruction errors for anomalou...Show MoreMetadata
Abstract:
Autoencoders (AEs) have been widely used for unsupervised anomaly detection. They learn from normal samples such that they produce high reconstruction errors for anomalous samples. However, AEs can exhibit the over-detection issue because they imperfectly reconstruct not only anomalous samples but also normal ones. To address this issue, we introduce an outlier-exposed style distillation network (OE-SDN) that mimics the mild distortions caused by an AE, which are termed as style translation. We use the difference between the outputs of the OE-SDN and AE as an alternative anomaly score. Experiments on anomaly classification and segmentation tasks show that the performance of our method is superior to existing methods.
The outlier-exposed style distillation network (OE-SDN) mimics the style translation and suppresses the content translation of the autoencoder (AE). Given an input image,...
Published in: IEEE Access ( Volume: 8)
Funding Agency:
Cognex Deep Learning Lab, Seoul, South Korea
Hwehee Chung received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), in 2016 and 2018, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His current research interest includes network debugging in the field of vision inspection systems in manufacturing.
Hwehee Chung received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), in 2016 and 2018, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His current research interest includes network debugging in the field of vision inspection systems in manufacturing.View more
Cognex Deep Learning Lab, Seoul, South Korea
Jongho Park received the B.S. and M.S. degrees in computer science and engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes optimizing the productivity in vision inspection industry by adapting deep learning algorithms.
Jongho Park received the B.S. and M.S. degrees in computer science and engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes optimizing the productivity in vision inspection industry by adapting deep learning algorithms.View more
Cognex Deep Learning Lab, Seoul, South Korea
Jongsoo Keum received the B.S. and M.S. degrees in electrical engineering and computer science from the Gwangju Institute of Science and Technology, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His main research interest includes optimized modeling of deep learning architecture in the field of vision inspection system in manufacturing.
Jongsoo Keum received the B.S. and M.S. degrees in electrical engineering and computer science from the Gwangju Institute of Science and Technology, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His main research interest includes optimized modeling of deep learning architecture in the field of vision inspection system in manufacturing.View more
Cognex Deep Learning Lab, Seoul, South Korea
Hongdo Ki received the B.S. and M.S. degrees in industrial engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes reducing the cost of applying a learning-based vision inspection system in manufacturing.
Hongdo Ki received the B.S. and M.S. degrees in industrial engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes reducing the cost of applying a learning-based vision inspection system in manufacturing.View more
Department of Industrial Engineering, Sungkyunkwan University, Suwon, South Korea
Seokho Kang received the B.S. and Ph.D. degrees in industrial engineering from Seoul National University, in 2011 and 2015, respectively. He was a Research Staff Member with the Samsung Advanced Institute of Technology. He is currently an Assistant Professor of systems management engineering (industrial engineering) with Sungkyunkwan University. His main research interest includes developing learning algorithms for effici...Show More
Seokho Kang received the B.S. and Ph.D. degrees in industrial engineering from Seoul National University, in 2011 and 2015, respectively. He was a Research Staff Member with the Samsung Advanced Institute of Technology. He is currently an Assistant Professor of systems management engineering (industrial engineering) with Sungkyunkwan University. His main research interest includes developing learning algorithms for effici...View more
Contents Cognex Deep Learning Lab, Seoul, South Korea
Hwehee Chung received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), in 2016 and 2018, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His current research interest includes network debugging in the field of vision inspection systems in manufacturing.
Hwehee Chung received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), in 2016 and 2018, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His current research interest includes network debugging in the field of vision inspection systems in manufacturing.View more
Cognex Deep Learning Lab, Seoul, South Korea
Jongho Park received the B.S. and M.S. degrees in computer science and engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes optimizing the productivity in vision inspection industry by adapting deep learning algorithms.
Jongho Park received the B.S. and M.S. degrees in computer science and engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes optimizing the productivity in vision inspection industry by adapting deep learning algorithms.View more
Cognex Deep Learning Lab, Seoul, South Korea
Jongsoo Keum received the B.S. and M.S. degrees in electrical engineering and computer science from the Gwangju Institute of Science and Technology, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His main research interest includes optimized modeling of deep learning architecture in the field of vision inspection system in manufacturing.
Jongsoo Keum received the B.S. and M.S. degrees in electrical engineering and computer science from the Gwangju Institute of Science and Technology, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His main research interest includes optimized modeling of deep learning architecture in the field of vision inspection system in manufacturing.View more
Cognex Deep Learning Lab, Seoul, South Korea
Hongdo Ki received the B.S. and M.S. degrees in industrial engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes reducing the cost of applying a learning-based vision inspection system in manufacturing.
Hongdo Ki received the B.S. and M.S. degrees in industrial engineering from Seoul National University, in 2015 and 2017, respectively. He is currently a Researcher with the Cognex Deep Learning Lab. His research interest includes reducing the cost of applying a learning-based vision inspection system in manufacturing.View more
Department of Industrial Engineering, Sungkyunkwan University, Suwon, South Korea
Seokho Kang received the B.S. and Ph.D. degrees in industrial engineering from Seoul National University, in 2011 and 2015, respectively. He was a Research Staff Member with the Samsung Advanced Institute of Technology. He is currently an Assistant Professor of systems management engineering (industrial engineering) with Sungkyunkwan University. His main research interest includes developing learning algorithms for efficient data-driven modeling and their applications to real-world data mining problems in manufacturing, healthcare, and materials informatics.
Seokho Kang received the B.S. and Ph.D. degrees in industrial engineering from Seoul National University, in 2011 and 2015, respectively. He was a Research Staff Member with the Samsung Advanced Institute of Technology. He is currently an Assistant Professor of systems management engineering (industrial engineering) with Sungkyunkwan University. His main research interest includes developing learning algorithms for efficient data-driven modeling and their applications to real-world data mining problems in manufacturing, healthcare, and materials informatics.View more