Abstract:

In order to improve the designing efficiency of the color style of the bedding items, the article uses K-means clustering algorithm and fuzzy membership function to calculate the dominant tone and color description to construct the color styles database of reference image for bedding items, to segment the foreground and background of source images and reference images. It also uses the expanded color harmony templates as the criteria and adopts standard Gauss distribution function to transfer the normality of foreground color in source images to the grey area in the color harmony template of reference images, thus to achieve the desired foreground color harmony transference. Then it adjusts the saturation and value of the merged images according to the variance and the mean value, and finally obtains target images whose styles are in accordance with the reference images, achieving the color harmony matching of the bedding items. The experimental results show the method in the article offers an effective way for color matching of the bedding items.

Key words: bed clothes, color harmony, color description semantics, color transfer