arrow
Volume 8, Issue 2
Fabric Color Difference Detection Based on SVM of Multi-dimension Features with Wavelet Kernel

Zhiyu Zhou, Rui Xu, Dichong Wu, Yingchun Liu & Zefei Zhu

Journal of Fiber Bioengineering & Informatics, 8 (2015), pp. 241-248.

Published online: 2015-08

Export citation
  • Abstract
Traditionally dyed fabric color difference detection is based on the image color characteristics in textile industry. However, relying solely on the single image color features can't effectively identify dyed fabric color difference with rich texture characteristics. In order to solve this problem, a new efficient color difference detection method based on multi-dimensional characteristics of Morlet Wavelet Kernel Support Vector Machine (MWSVM) is proposed in this paper. Firstly the dyed fabric image to be detected is divided into some appropriate sub-blocks in the LAB color space. The LAB histograms of the image in those sub-blocks are extracted. In addition, the Local Binary Pattern (LBP) algorithm is applied to extract the image texture features in those different divided regions. Then the Grey Relational Grade (GRG) between the sample image and the detected image is calculated. Finally the LAB histograms, the LBP features and the GRG are used as the input image data for the MWSVM algorithm to detect color difference of dyed fabrics. The experimental results show that the proposed method can detect dyed fabric color difference more efficiently and accurately. The classification accuracy rate as high as 87.5%.
  • AMS Subject Headings

  • Copyright

COPYRIGHT: © Global Science Press

  • Email address
  • BibTex
  • RIS
  • TXT
@Article{JFBI-8-241, author = {}, title = {Fabric Color Difference Detection Based on SVM of Multi-dimension Features with Wavelet Kernel}, journal = {Journal of Fiber Bioengineering and Informatics}, year = {2015}, volume = {8}, number = {2}, pages = {241--248}, abstract = {Traditionally dyed fabric color difference detection is based on the image color characteristics in textile industry. However, relying solely on the single image color features can't effectively identify dyed fabric color difference with rich texture characteristics. In order to solve this problem, a new efficient color difference detection method based on multi-dimensional characteristics of Morlet Wavelet Kernel Support Vector Machine (MWSVM) is proposed in this paper. Firstly the dyed fabric image to be detected is divided into some appropriate sub-blocks in the LAB color space. The LAB histograms of the image in those sub-blocks are extracted. In addition, the Local Binary Pattern (LBP) algorithm is applied to extract the image texture features in those different divided regions. Then the Grey Relational Grade (GRG) between the sample image and the detected image is calculated. Finally the LAB histograms, the LBP features and the GRG are used as the input image data for the MWSVM algorithm to detect color difference of dyed fabrics. The experimental results show that the proposed method can detect dyed fabric color difference more efficiently and accurately. The classification accuracy rate as high as 87.5%.}, issn = {2617-8699}, doi = {https://doi.org/10.3993/jfbim00108}, url = {http://global-sci.org/intro/article_detail/jfbi/4703.html} }
TY - JOUR T1 - Fabric Color Difference Detection Based on SVM of Multi-dimension Features with Wavelet Kernel JO - Journal of Fiber Bioengineering and Informatics VL - 2 SP - 241 EP - 248 PY - 2015 DA - 2015/08 SN - 8 DO - http://doi.org/10.3993/jfbim00108 UR - https://global-sci.org/intro/article_detail/jfbi/4703.html KW - LAB Color Space KW - LBP KW - Grey Relational Grade (GRG) KW - SVM KW - Morlet Wavelet Kernel KW - Color Difference Detection AB - Traditionally dyed fabric color difference detection is based on the image color characteristics in textile industry. However, relying solely on the single image color features can't effectively identify dyed fabric color difference with rich texture characteristics. In order to solve this problem, a new efficient color difference detection method based on multi-dimensional characteristics of Morlet Wavelet Kernel Support Vector Machine (MWSVM) is proposed in this paper. Firstly the dyed fabric image to be detected is divided into some appropriate sub-blocks in the LAB color space. The LAB histograms of the image in those sub-blocks are extracted. In addition, the Local Binary Pattern (LBP) algorithm is applied to extract the image texture features in those different divided regions. Then the Grey Relational Grade (GRG) between the sample image and the detected image is calculated. Finally the LAB histograms, the LBP features and the GRG are used as the input image data for the MWSVM algorithm to detect color difference of dyed fabrics. The experimental results show that the proposed method can detect dyed fabric color difference more efficiently and accurately. The classification accuracy rate as high as 87.5%.
Zhiyu Zhou, Rui Xu, Dichong Wu, Yingchun Liu & Zefei Zhu. (2019). Fabric Color Difference Detection Based on SVM of Multi-dimension Features with Wavelet Kernel. Journal of Fiber Bioengineering and Informatics. 8 (2). 241-248. doi:10.3993/jfbim00108
Copy to clipboard
The citation has been copied to your clipboard