Defect detection method using deep convolutional neural network, support vector machine and template matching techniques

Fusaomi Nagata, Kenta Tokuno, Kazuki Mitarai, Akimasa Otsuka, Takeshi Ikeda, Hiroaki Ochi, Keigo Watanabe, Maki K. Habib

Research output: Contribution to journal › Article

Abstract

In this paper, a defect detection method using deep convolutional neural network (DCNN), support vector machine (SVM) and template matching techniques is introduced. First, a DCNN for visual inspection is designed and trained using a large number of images to inspect undesirable defects such as crack, burr, protrusion, chipping, spot and fracture phenomena which appear in the manufacturing process of resin molded articles. Then the trained DCNN named sssNet and well-known AlexNet are, respectively, incorporated with two SVMs to classify sample images with high recognition rate into accept as OK category or reject as NG one, in which compressed feature vectors obtained from the DCNNs are used as inputs for the SVMs. The performances of the two types of SVMs with the DCNNs are compared and evaluated through training and classification experiments. Finally, a template matching technique is further proposed to efficiently extract important target areas from original training and test images. This will be able to enhance the reliability and accuracy for defect detection.

Original languageEnglish
JournalArtificial Life and Robotics
DOIs
Publication statusPublished - Jan 1 2019

Fingerprint

Template matching
Support vector machines
Neural networks
Resins
Inspection
Cracks
Defects
Defect detection
Support Vector Machine
Experiments

Keywords

  • Deep convolutional neural network (DCNN)
  • Defect detection system
  • Support vector machine (SVM)
  • Template matching

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Artificial Intelligence

Cite this

Defect detection method using deep convolutional neural network, support vector machine and template matching techniques. / Nagata, Fusaomi; Tokuno, Kenta; Mitarai, Kazuki; Otsuka, Akimasa; Ikeda, Takeshi; Ochi, Hiroaki; Watanabe, Keigo; Habib, Maki K.

In: Artificial Life and Robotics, 01.01.2019.

Research output: Contribution to journal › Article

Nagata, F, Tokuno, K, Mitarai, K, Otsuka, A, Ikeda, T, Ochi, H, Watanabe, K & Habib, MK 2019, 'Defect detection method using deep convolutional neural network, support vector machine and template matching techniques', Artificial Life and Robotics. https://doi.org/10.1007/s10015-019-00545-x
Nagata F, Tokuno K, Mitarai K, Otsuka A, Ikeda T, Ochi H et al. Defect detection method using deep convolutional neural network, support vector machine and template matching techniques. Artificial Life and Robotics. 2019 Jan 1. https://doi.org/10.1007/s10015-019-00545-x
Nagata, Fusaomi ; Tokuno, Kenta ; Mitarai, Kazuki ; Otsuka, Akimasa ; Ikeda, Takeshi ; Ochi, Hiroaki ; Watanabe, Keigo ; Habib, Maki K. / Defect detection method using deep convolutional neural network, support vector machine and template matching techniques. In: Artificial Life and Robotics. 2019.
@article{1981103c5c584b0d86b0431429f49cb0,
title = "Defect detection method using deep convolutional neural network, support vector machine and template matching techniques",
abstract = "In this paper, a defect detection method using deep convolutional neural network (DCNN), support vector machine (SVM) and template matching techniques is introduced. First, a DCNN for visual inspection is designed and trained using a large number of images to inspect undesirable defects such as crack, burr, protrusion, chipping, spot and fracture phenomena which appear in the manufacturing process of resin molded articles. Then the trained DCNN named sssNet and well-known AlexNet are, respectively, incorporated with two SVMs to classify sample images with high recognition rate into accept as OK category or reject as NG one, in which compressed feature vectors obtained from the DCNNs are used as inputs for the SVMs. The performances of the two types of SVMs with the DCNNs are compared and evaluated through training and classification experiments. Finally, a template matching technique is further proposed to efficiently extract important target areas from original training and test images. This will be able to enhance the reliability and accuracy for defect detection.",
keywords = "Deep convolutional neural network (DCNN), Defect detection system, Support vector machine (SVM), Template matching",
author = "Fusaomi Nagata and Kenta Tokuno and Kazuki Mitarai and Akimasa Otsuka and Takeshi Ikeda and Hiroaki Ochi and Keigo Watanabe and Habib, {Maki K.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/s10015-019-00545-x",
language = "English",
journal = "Artificial Life and Robotics",
issn = "1433-5298",
publisher = "Springer Japan",

}

TY - JOUR

T1 - Defect detection method using deep convolutional neural network, support vector machine and template matching techniques

AU - Nagata, Fusaomi

AU - Tokuno, Kenta

AU - Mitarai, Kazuki

AU - Otsuka, Akimasa

AU - Ikeda, Takeshi

AU - Ochi, Hiroaki

AU - Watanabe, Keigo

AU - Habib, Maki K.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In this paper, a defect detection method using deep convolutional neural network (DCNN), support vector machine (SVM) and template matching techniques is introduced. First, a DCNN for visual inspection is designed and trained using a large number of images to inspect undesirable defects such as crack, burr, protrusion, chipping, spot and fracture phenomena which appear in the manufacturing process of resin molded articles. Then the trained DCNN named sssNet and well-known AlexNet are, respectively, incorporated with two SVMs to classify sample images with high recognition rate into accept as OK category or reject as NG one, in which compressed feature vectors obtained from the DCNNs are used as inputs for the SVMs. The performances of the two types of SVMs with the DCNNs are compared and evaluated through training and classification experiments. Finally, a template matching technique is further proposed to efficiently extract important target areas from original training and test images. This will be able to enhance the reliability and accuracy for defect detection.

AB - In this paper, a defect detection method using deep convolutional neural network (DCNN), support vector machine (SVM) and template matching techniques is introduced. First, a DCNN for visual inspection is designed and trained using a large number of images to inspect undesirable defects such as crack, burr, protrusion, chipping, spot and fracture phenomena which appear in the manufacturing process of resin molded articles. Then the trained DCNN named sssNet and well-known AlexNet are, respectively, incorporated with two SVMs to classify sample images with high recognition rate into accept as OK category or reject as NG one, in which compressed feature vectors obtained from the DCNNs are used as inputs for the SVMs. The performances of the two types of SVMs with the DCNNs are compared and evaluated through training and classification experiments. Finally, a template matching technique is further proposed to efficiently extract important target areas from original training and test images. This will be able to enhance the reliability and accuracy for defect detection.

KW - Deep convolutional neural network (DCNN)

KW - Defect detection system

KW - Support vector machine (SVM)

KW - Template matching

UR - http://www.scopus.com/inward/record.url?scp=85068797129&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068797129&partnerID=8YFLogxK

U2 - 10.1007/s10015-019-00545-x

DO - 10.1007/s10015-019-00545-x

M3 - Article

JO - Artificial Life and Robotics

JF - Artificial Life and Robotics

SN - 1433-5298

ER -

Access to Document