A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects

Gwang Myong Go; Seok Jun Bu; Sung Bae Cho

doi:10.1007/978-3-030-62365-4_46

A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects

Gwang Myong Go, Seok Jun Bu, Sung Bae Cho

College of Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

For defect inspection using computer vision, deep learning models have been introduced to improve the conventional rule-based pattern analysis. A lot of data is a prerequisite to the success of them, but the on-the-spot industrial field suffers from lack of data. In this paper, we propose a deep metric neural network to improve the performance even with insufficient data imbalanced in class. The model is verified with the dataset of new products by evaluating the accuracy with 10-fold cross-validation. Our model is based on the data in the smallest category, 1.2 K, which achieves the highest performance of 90.42% using sampled pairs without using all the data for training. High accuracy has been achieved and proven applicability in the industry compared to the conventional machine learning models.

Original language	English
Title of host publication	Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings
Editors	Cesar Analide, Paulo Novais, David Camacho, Hujun Yin
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	485-494
Number of pages	10
ISBN (Print)	9783030623647
DOIs	https://doi.org/10.1007/978-3-030-62365-4_46
Publication status	Published - 2020
Event	21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020 - Guimaraes, Portugal Duration: 2020 Nov 4 → 2020 Nov 6

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12490 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020
Country/Territory	Portugal
City	Guimaraes
Period	20/11/4 → 20/11/6

Bibliographical note

Publisher Copyright:
© 2020, Springer Nature Switzerland AG.

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-62365-4_46

Cite this

Go, G. M., Bu, S. J., & Cho, S. B. (2020). A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects. In C. Analide, P. Novais, D. Camacho, & H. Yin (Eds.), Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings (pp. 485-494). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12490 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-62365-4_46

Go, Gwang Myong ; Bu, Seok Jun ; Cho, Sung Bae. / A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects. Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings. editor / Cesar Analide ; Paulo Novais ; David Camacho ; Hujun Yin. Springer Science and Business Media Deutschland GmbH, 2020. pp. 485-494 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{9cc9059371d547f1b408c79bd1434edc,

title = "A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects",

abstract = "For defect inspection using computer vision, deep learning models have been introduced to improve the conventional rule-based pattern analysis. A lot of data is a prerequisite to the success of them, but the on-the-spot industrial field suffers from lack of data. In this paper, we propose a deep metric neural network to improve the performance even with insufficient data imbalanced in class. The model is verified with the dataset of new products by evaluating the accuracy with 10-fold cross-validation. Our model is based on the data in the smallest category, 1.2 K, which achieves the highest performance of 90.42% using sampled pairs without using all the data for training. High accuracy has been achieved and proven applicability in the industry compared to the conventional machine learning models.",

author = "Go, {Gwang Myong} and Bu, {Seok Jun} and Cho, {Sung Bae}",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; 21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020 ; Conference date: 04-11-2020 Through 06-11-2020",

year = "2020",

doi = "10.1007/978-3-030-62365-4_46",

language = "English",

isbn = "9783030623647",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "485--494",

editor = "Cesar Analide and Paulo Novais and David Camacho and Hujun Yin",

booktitle = "Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings",

address = "Germany",

}

Go, GM, Bu, SJ & Cho, SB 2020, A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects. in C Analide, P Novais, D Camacho & H Yin (eds), Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12490 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 485-494, 21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020, Guimaraes, Portugal, 20/11/4. https://doi.org/10.1007/978-3-030-62365-4_46

A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects. / Go, Gwang Myong; Bu, Seok Jun; Cho, Sung Bae.
Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings. ed. / Cesar Analide; Paulo Novais; David Camacho; Hujun Yin. Springer Science and Business Media Deutschland GmbH, 2020. p. 485-494 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12490 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects

AU - Go, Gwang Myong

AU - Bu, Seok Jun

AU - Cho, Sung Bae

PY - 2020

Y1 - 2020

N2 - For defect inspection using computer vision, deep learning models have been introduced to improve the conventional rule-based pattern analysis. A lot of data is a prerequisite to the success of them, but the on-the-spot industrial field suffers from lack of data. In this paper, we propose a deep metric neural network to improve the performance even with insufficient data imbalanced in class. The model is verified with the dataset of new products by evaluating the accuracy with 10-fold cross-validation. Our model is based on the data in the smallest category, 1.2 K, which achieves the highest performance of 90.42% using sampled pairs without using all the data for training. High accuracy has been achieved and proven applicability in the industry compared to the conventional machine learning models.

AB - For defect inspection using computer vision, deep learning models have been introduced to improve the conventional rule-based pattern analysis. A lot of data is a prerequisite to the success of them, but the on-the-spot industrial field suffers from lack of data. In this paper, we propose a deep metric neural network to improve the performance even with insufficient data imbalanced in class. The model is verified with the dataset of new products by evaluating the accuracy with 10-fold cross-validation. Our model is based on the data in the smallest category, 1.2 K, which achieves the highest performance of 90.42% using sampled pairs without using all the data for training. High accuracy has been achieved and proven applicability in the industry compared to the conventional machine learning models.

UR - http://www.scopus.com/inward/record.url?scp=85097138942&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85097138942&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-62365-4_46

DO - 10.1007/978-3-030-62365-4_46

M3 - Conference contribution

AN - SCOPUS:85097138942

SN - 9783030623647

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 485

EP - 494

BT - Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings

A2 - Analide, Cesar

A2 - Novais, Paulo

A2 - Camacho, David

A2 - Yin, Hujun

PB - Springer Science and Business Media Deutschland GmbH

T2 - 21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020

Y2 - 4 November 2020 through 6 November 2020

ER -

Go GM, Bu SJ, Cho SB. A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects. In Analide C, Novais P, Camacho D, Yin H, editors, Intelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 485-494. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-62365-4_46

A Deep Metric Neural Network with Disentangled Representation for Detecting Smartphone Glass Defects

Abstract

Publication series

Conference

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this