Home // PATTERNS 2017, The Ninth International Conferences on Pervasive Patterns and Applications // View article

Detection and Classification of Defective Electronic Circuit Boards Using CNN Features and Keypoint Extraction

Authors:
Yohei Takada
Tokiko Shiina
Hiroyasu Usami
Yuji Iwahori
Bhuyan Manas Kamal

Keywords: Defect Detection; Defect Classification; CNN; SVM; SURF.

Abstract:
This paper proposes a method for defect detection and classification of electronic circuit board by extracting keypoints without reference images. The final purpose is to distinguish a problematic defect, such as disconnection from a non-defect, dust in the manufacturing process et al. Keypoints is extracted from the electronic circuit board image, then a patch image is cropped using obtained keypoint information, such as the position. The cropped images are used as input to CNN (Convolutional Neural Network) and 4096-dimensional features are obtained in the final layer of the full connected layers. SVM (Support Vector Machine) is introduced for learning and classification using CNN features. The effectiveness of the proposed method is confirmed through a detection experiment using actual electronic circuit board images containing defects and by comparing the results with the previous method.

Pages: 113 to 116

Copyright: Copyright (c) IARIA, 2017

Publication date: February 19, 2017

Published in: conference

ISSN: 2308-3557

ISBN: 978-1-61208-534-0

Location: Athens, Greece

Dates: from February 19, 2017 to February 23, 2017