Repeat pattern segmentation of print fabric based on adaptive template matching

Zhong Xiang,Yang Liu,Zhenyu Wu,Xudong Hu,Ding Zhou,Miao Qian,Miao Ma

doi:10.1177/1558925020973285

Zhong Xiang, Yang Liu + Show 5 more

Open Access

https://doi.org/10.1177/1558925020973285

Copy DOI

Abstract

Patterned fabrics are generally constructed from the periodic repetition of a primitive pattern unit. Repeat pattern segmentation of printed fabrics has a very significant impact on the pattern retrieval and pattern defect detection. In this paper, we propose a new approach for repeat pattern segmentation by employing the adaptive template matching method. In contrast to the traditional method for template matching, the proposed algorithm first selects an adaptive size template image in the repeat pattern image based on the size of the original image and its local maximum edge density. Then it uses the sum of absolute differences as the matching features to identify the matched regions in the original image, and the minimum envelope border of the primitive pattern, typically as a parallelogram, can be determined from the results of the four adjacent matched templates. Finally, image traversal base on the obtained parallelogram is implemented over the original image using minimum information loss theory to produce a well-segmented primitive pattern with a complete edge structure. The results from the experiments conducted using an extensive database of real fabric images show that the proposed algorithm has the advantage of rotation invariance and scaling invariance and will not be affected when the background or foreground color is changed.

Highlights

Patterned fabrics are generally constructed from periodic primitive pattern, which is their most typical characteristic
Adaptive template matching is used for periodic pattern size estimation and repeat pattern primitive segmentation
Unlike traditional template matching algorithms, the proposed adaptive template matching algorithm selects a part of the original image as the template image by analyzing the characteristics of the patterned fabric using the maximum edge density principle

Summary

Introduction

Patterned fabrics are generally constructed from periodic primitive pattern, which is their most typical characteristic. To reduce the computation and memory requirements, Unser proposed the sum and difference histograms (SDH) method as an alternative to the GLCM for conducting the texture analysis.[8] SDH describes the probability distribution of the sum and difference between adjacent pixels at a certain distance It reflects the complex information of the image regarding the direction, distance, and degree of variability when the distribution of the patterns or textures shows some regularities.[9] Lizarraga-Morales et al uses the homogeneity computed from SDH to extract the texture periodicity corresponding to a square or rectangular unit.[10] In his further research, this method was used to detect defects in fabrics with regular textures, and the experiment results demonstrated that it was superior to other state-ofthe-art algorithms.[11] an obvious weakness of this method is the lack of color information, so that it cannot be applied to colorful images. That we can reduce the computation time by limiting the traversing region For this purpose, we stored the edge density value and the center coordinate value of any effective window in the DTI retrieval process, but only the windows meets equation (4) are selected for template matching calculation while the others are skipped:.

Findings

Experiments and discussion

Conclusion