Texture Segmentation and Contour Extraction of Warp Knitting Jacquard Fabric Image

Abstract:

To improve the accuracy and efficiency of warp knitting CAD, a new approach to texture segmentation and contour extraction based on wavelet-transform, K-means clustering and Canny operator is proposed in this paper. The procedure is described as follows. Firstly the Daubechies wavelet and pyramid-structure is selected, then the approach decomposes the low frequency part of the fabric image. Secondly starting from the highest scaling level and considering all the four sub-bands, the image is automatic clustered and segmented by K-means clustering. During the process we should calculate the average mean and standard deviation of the three high-frequency coefficients at the current level and transform the higher scaling level label values to have the same mean and standard deviation. Lastly the texture’s contour of image is traced and extracted by Canny operator. The results show that this approach is a feasible way for jacquard fabric.

• Periodical: Applied Mechanics and Materials (Volumes 201-202)
• Pages: 1203-1206
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.201-202.1203
• Online since: October 2012
• Authors: Yang Zhang, You Cheng Tong, Jun Zhou Yao
• Keywords: Canny Operator, Contour Extraction, Jacquard Fabric, K-Means Clustering, Texture Segmentation, Wavelet Transform (WT)

References

[1] D. Dunn, W.E. Higgins, J. Wakeley: IEEE Transactions on PAMI, (1994) No.2, pp.130-149.

[2] G.H. Wu, Y.J. Zhang, X.G. Proc. ICIP'99 (USA:IEEE Publisher,1999), pp.435-439.

[3] G.H. WU, Y.J Zhang, X.G. Lin: Journal of Image and Graphics, (2001) No.4, pp.333-337.