基于自适应模板匹配的四方连续花型最小单元提取

doi:10.19398/j.att.202103018

现代纺织技术 ›› 2022, Vol. 30 ›› Issue (2): 191-196.DOI: 10.19398/j.att.202103018

基于自适应模板匹配的四方连续花型最小单元提取

龙颖^1a, 吕叶馨^1a^,², 郭子翊^1a, 黄振华^1a, 邹奉元^1a^,^1b^,^1c

1.浙江理工大学,a.服装学院,b.丝绸文化传承与产品设计数字化技术文化和旅游部重点实验室,c.浙江省服装工程技术研究中心,杭州　310018
2.金华职业技术学院艺术创意学院,浙江金华 321000

收稿日期:2021-03-11 出版日期:2022-03-10 网络出版日期:2021-07-08
作者简介:龙颖(1996-),男,湖南岳阳人,硕士研究生,主要从事服装数字化技术方面的研究。
基金资助:
浙江省教育厅项目(Y201942686)

Extraction of minimum units of four-consecutive patternbased on adaptive template matching

LONG Ying^1a, LÜ Yexin^1a^,², GUO Ziyi^1a, HUANG Zhenhua^1a, ZOU Fengyuan^1a^,^1b^,^1c

1a. School of Fashion Design & Engineering; 1b. Key Laboratory of Silk Culture Heritage andProducts Design Digital Technology, Ministry of Culture and Tourism; 1c. Zhejiang Provincial ResearchCenter of Clothing Engineering Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China
2. College of Creative Arts, Jinhua Polytechnic, Jinhua 321000, China

Received:2021-03-11 Published:2022-03-10 Online:2021-07-08

摘要/Abstract

摘要：

为实现连缀式及重叠式四方连续花型最小单元的自动提取,对矩形模板边长占比进行迭代,探究了矩形模板大小对全局模板匹配的影响,再利用垂直矢量法分析了匹配所得矩形的中心点点集,揭示出像元重复规律与最小单元的对应关系;以相似度均值度量提取结果的准确性,阐明了最小单元重复次数与相似度均值的关系;通过分析模板匹配的适用性,提出了自适应阈值模板匹配的方法,改进后相似度均值平均可达94.29%。结果表明,该方法能有效实现连缀式及重叠式四方连续花型最小单元的提取。

关键词: 最小单元提取, 自适应模板匹配, 四方连续花型, 相似性匹配, 直方图相似度

Abstract:

In order to realize the automatic extraction of minimum units of concatenated and overlapped four-consecutive patterns, this article iterated on the ratio of side length of rectangular template and investigated its impact on global template matching. The vertical vector method was adopted to analyze the center point set of the rectangle obtained from matching. The correspondence between the repetition law of pixels and the minimum unit was revealed. The accuracy of extraction results was measured by similarity mean, and the relationship between repetition times of minimum unit and similarity mean was clarified. By analyzing the applicability of template matching, an adaptive threshold template matching method was put forward. The average similarity can reach 94.29% after improvement. The results show that this method can effectively achieve the extraction of minimum units of concatenated and overlapped four-consecutive patterns.

Key words: extraction of minimum units, adaptive template matching, four-consecutive pattern, similarity matching, histogram similarity

中图分类号:

TS941.26

龙颖, 吕叶馨, 郭子翊, 黄振华, 邹奉元. 基于自适应模板匹配的四方连续花型最小单元提取[J]. 现代纺织技术, 2022, 30(2): 191-196.

LONG Ying, LÜ Yexin, GUO Ziyi, HUANG Zhenhua, ZOU Fengyuan. Extraction of minimum units of four-consecutive patternbased on adaptive template matching[J]. Advanced Textile Technology, 2022, 30(2): 191-196.

图/表 12

图1 四方连续纹样

Fig.1 The organization form of four-consecutive pattern

图2 单一元素重复规律示意

Fig.2 Schematic diagram of repetition law of single element

图3 垂直矢量法

Fig.3 Vertical vector method

图4 拼合部分(中)与原图(右)对应位置直方图相似度

Fig.4 Histogram similarity between the split part and the corresponding position of the original image

图5 不同模板边长占比的相似度均值

Fig.5 Average similarity of the ratio ofside length of different templates

表1 模板边长最佳占比区间

Tab.1 The optimum proportion of template side length

指标	样本数量
指标	6	9	12	15
最佳占比区间	0.28~0.32	0.25~0.35	0.21~0.38	0.21~0.38
异常数	0	0	1	2

图6 3类花型样本

Fig.6 Three types of pattern samples

表2 相似度均值计算结果

Tab.2 Thecalculated result ofsimilarity mean value

样本类别	样本数量	相似度均值/%
Ⅰ类	17	88.38
Ⅱ类	9	94.41
Ⅲ类	14	92.86
平均	-	91.88

图7 不同重复次数的四方连续纹样

Fig.7 Four-continuous pattern with different numbers of minimum units

图8 不同阈值下的相似度均值

Fig.8 The mean value of similarity under different thresholds

表3 自适应阈值前后相似度均值对比

Tab.3 Comparison of similarity mean value of adaptive threshold before and after

样本类别	样本数量	平均相似度均值(前)/%	平均相似度均值(后)/%	变化水平/%
Ⅰ类	17	88.38	90.72	2.34
Ⅱ类	9	94.41	96.32	1.91
Ⅲ类	14	92.86	95.83	2.97
平均值	-	91.88	94.29	2.41

图9 验证实验样本及提取结果

Fig.9 Validation experiment samples and extraction results

参考文献 12

[1]	陶晨, 周赳, 奚柏君. 织物图案连续性识别[J]. 纺织学报, 2016, 37(8):37-40.
	TAO Chen, ZHOU Jiu, XI Bojun. Continuity identification of fabric patterns[J]. Journal of Textile Research, 2016, 37(8):37-40.
[2]	ZHANG J, PAN R R, GAO W D, et al. Automatic recognition of the color effect of yarn-dyed fabric by the smallest repeat unit recognition algorithm[J]. Textile Research Journal, 2015, 85(4):432-446. DOI URL
[3]	XIAO Z T, GUO Y M, GENG L, et al. Automatic recognition of woven fabric pattern based on TILT[J/OB]. Mathematical Problems in Engineering, 2018: 9707104. https://doi.org/10.1155/2018/9707104.
[4]	ZHENG D J, WANG L H. Multi-scale density detection for yarn-dyed fabrics with deformed repeat patterns[J]. Textile Research Journal, 2017, 87(20):2524-2540. DOI URL
[5]	TAO C, ZHOU J, YIN M F. Automatic identification of textile pattern consecutiveness based on similarity space[J]. Textile Research Journal, 2017, 87(2):224-231. DOI URL
[6]	景军锋, 杨盼盼, 李鹏飞. 采用距离匹配函数的印花织物图案周期测定[J]. 纺织学报, 2015, 36(12):98-103.
	JING Junfeng, YANG Panpan, LI Pengfei. Determination on design cycle of printed fabrics based on distance matching function[J]. Journal of Textile Research, 2015, 36(12):98-103.
[7]	KUO J, SHIN C Y, LEE J. Repeat pattern segmentation of printed fabrics by Hough transform method[J]. Textile Research Journal, 2005, 75(11):779-783. DOI URL
[8]	KUO J, LEE C L, SHIH C Y. Image database of printed fabric with repeating dot patterns part (I)-image archiving[J]. Textile Research Journal, 2017, 87(17):2089-2105. DOI URL
[9]	刘兆辉, 张华熊, 林翔宇. 基于内容特征的四方连续纹样自动生成[J]. 计算机系统应用, 2019, 28(5):156-160.
	LIU Zhaohui, ZHANG Huaxiong, LIN Xiangyu. Automatic generation of quadrangle continuous patterns based on content characteristics[J]. Computer Systems & Applications, 2019, 28(5):156-160.
[10]	赵启. 图像匹配算法研究[D]. 西安:西安电子科技大学, 2013.
	ZHAO Qi. Research on Algorithm of Image Matching[D]. Xi'an: Xidian University, 2013.
[11]	OTSU N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1):62-66. DOI URL
[12]	汪会, 孙洁, 丁笑君, 等. 织物纹样特征提取与匹配方法比较[J]. 纺织学报, 2020, 41(4):45-50.
	WANG Hui, SUN Jie, DING Xiaojun, et al. Comparison of feature extracting and matching methods for fabric patterns[J]. Journal of Textile Research, 2020, 41(4):45-50.

基于自适应模板匹配的四方连续花型最小单元提取

Extraction of minimum units of four-consecutive patternbased on adaptive template matching

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 12

相关文章 0

编辑推荐

Metrics