Abstract:

Manual extraction of embroidered patterns with human vision is time consuming, lab-intensive and subjectivitve. To overcome these difficulties, image analysis was adopted to complete the automatic recognition and aesthetic analysis of the embroidered patterns on the traditional consumes fabrics. Digital camera was used to realize the image acquisition of embroidered fabric. Image smoothing was then applied to the fabric image to remove the noise information and the fabric image was converted to Lab color space from RGB color space. K-Means clustering method was adopted to realize the color clustering based on a and b components in Lab color space. The accurate pattern and color information of the embroidered fabric were obtained based on the clustering results. The research results show that the image analysis can achieve the automatic extraction of embroidered patterns and the new segmentation method can be applied to quite a number of embroidered fabric patterns, which provides a brand new way for sample collection of aesthetic research.

Key words: traditional construme, embroidered pattern, image analysis, clustering analysis, color value analysis