Computerized color distinguishing system for color printed fabric by using the approach of probabilistic neural network

This article proposes an innovative color printed fabric computer color distinguishing system whose main functions are to precisely distinguish the printed fabric pattern colors and match colors to improve the current time-consuming color distinguishing conducted by manpower. The RGB color mode is an industrial color standard, by which the change and overlapping of color channels of red, green, and blue represent types of colors. RGB stands for red, green, and blue, respectively. It is one of the most widely used color system and covers almost all of the colors sensible to human vision. Hence, this paper adopts the RGB color mode to present color printed fabric images. First, to reduce the color distinguishing computation, a genetic algorithm was applied in search of small images of the same color in the original color printed fabric. Then, color distinguishing computation was conducted by a probabilistic neural network (PNN), which has the advantage of a very fast learning speed. Finally, PANTONE(R) standard color tickets were applied in matching colors. The experimental results revealed that the PNN design can easily realize and achieve accurate, fast color classification. It is proved that this color distinguishing system can be practically applied in printed fabric color distinguishing and matching.