A Method of Pattern Feature Extraction for Clothing Texture

摘要

In order to improve the information elements in fashion design display ability, feature extraction and pattern information for modeling clothing texture, a clothing texture pattern texture segmentation method the dynamic characteristics and speed up robust feature matching based on sparse interference on clothing texture acquisition case scattered point segmentation and denoising, color the introduction of Hessian component RGB scale clothing texture pattern decomposition method to decompose, clothing texture pattern feature extraction by the interference of texture segmentation, using Taubin smoothing operator clothing texture map in case of three-dimensional reconstruction of image pixel space, linear texture pattern reorganization of the clothing RGB component by Radon transform, to achieve the extraction of pattern features for clothing texture. The simulation results show that the extracted pattern features by using this method to achieve the texture of the clothing, clothing color information fusion of shape vision information can improve the information expression ability of dress texture, and the feature extraction of output pattern is better.