A study of automated color, shape and texture analysis of Tatami embroidery fabrics

Abstract

This paper proposes an automated analysis system for Tatami embroidery fabric images that automates color analysis, pattern shape analysis and texture analysis. Firstly, the RGB (red, green and blue) image of embroidery fabric is obtained by a color scanner. In color analysis, the wavelet transforms and median filter are used for RGB image preprocessing, and then the Fuzzy C-Means (FCM) clustering method is used for the binary region splitting method, the color features of statistical values of colors and number of colors can be obtained. In the pattern shape analysis, the individual pattern components are segmented by the separated colors, and the shape features of pattern components are obtained by moment invariants. Finally, in the texture analysis of Tatami embroidery fabrics, the gray image of the weave block is segmented from the pattern component region, after the entropy filter, the parallel division lines are segmented by a threshold value, and then the angle and spacing between Tatami embroidery parallel division lines are obtained by Hough transform to represent texture features. The experimental results show that this method can implement Tatami embroidery fabric color, shape and texture analyses automatically.