Abstract—This paper introduces a new fabric segmentation approach for detecting fabric defects using auto-correlation function. This proposed approach consists of 4 steps: 1) calculating the texture primitive template by auto-correlation function from defect free fabric image in train phase, 2) enhancing the defect areas, through calculation the difference between each texture primitive template and texture image, 3) constructing the mean image to reduce high frequent information of background image, and 4) compute a perfect automatic threshold to present a binary image as a defect pattern. At the end of paper, validity and robustness of the new approach were proved by some experiments done on different defect types. The results indicate that proposed method is implementable on both patterned and unpatterned fabrics.
At the end of paper, validity and robustness of the new approach were proved by some experiments done on different defect types. The results indicate that proposed method is implementable on both patterned and unpatterned fabrics.
Index Terms—Texture primitive template, defect pattern, image enhancement, defect segmentation.
The authors are with the Department of Computer sciences and Engineering, Shiraz University, Iran (e-mail: ehoseini@cse.shirazu.ac.irfarnoush.farhadi@ gmail.com, tajeri@ shirazu.ac.ir).
Cite: Elham Hoseini, Farnoush Farhadi, and Farshad Tajeripour, "Fabric Defect Detection Using Auto-Correlation Function," International Journal of Computer Theory and Engineering vol. 5, no. 1, pp. 114-117, 2013.
Copyright © 2008-2024. International Association of Computer Science and Information Technology. All rights reserved.
