Color representation model for pre-colored fiber blends based on local-to-global features fusion

Abstract

This paper proposes a color representation model of pre-colored fiber blends that combines global and local features, and the main purpose of this model is to establish scientific and effective evaluation standards for the colorimetry index of colored spun fabric, which can be used to solve the difficulty faced by people to accurately describe the overall coloration characteristics of dyed fibers in colored spinning process. Firstly, the third-order color moment features and three structure descriptor features are extracted, and decision level image fusion of feature is performed to represent the color of the samples. Secondly, the normalized difference of the fusing features between different samples are obtained, and it makes a correlation analysis of the difference of the mass ratio of dyed fiber. The experimental results show that the color representation model can effectively represent the change of the mass ratio of dyed fibers, and meanwhile it is more practical compared with the traditional spectrophotometry.