Woven Fabric Image Reconstruction Based on Dictionary Learning

摘要：

　　Texture is an important feature for woven fabric.Reconstruction of woven fabric image for woven fabric texture information.Woven fabric original image divided into small patches as training samples, choose the K-SVD algorithm to obtain optimal dictionary and coefficients from training samples.The reconstructed image by a linear combination of optimal dictionary atoms and coefficients.With SSIM as the evaluation index of the image to evaluate the quality of the reconstructed image.The result show that the dictionary learning method can reconstruct the woven fabric image very well, and for different woven fabric structures, the results will be different.

关键词： woven fabric texture image dictionary learning image reconstruction

作者: Jie Wang Jiang Wang Liqing Li Jun Wang

作者单位: Collage of Textiles, Donghua University, China Collage of Textiles, Donghua University, China;Key Laboratory of Textile Science & Technology, Minis

会议类型: 国际会议

会议名称: 2017 Sino-Africa International Symposium on Textiles and Apparel (SAISTA2017)(2017中非纺织服装国际论坛)

会议地点: 上海

会议语种:英文

页码: 250-253

在线出版日期: 2017-11-18（万方平台首次上网日期，不代表论文的发表时间）