A color transfer model for colored spun fabrics based on monochromatic fibers

Abstract

The structure differences between fibers and fabrics result in noticeable color variations, posing challenges in the actual production of colored spun fabrics. To address this issue, we studied the color relationship between the fibers and fabrics by leveraging reflectance and introduced a novel color transfer model to describe the accurate color changing trends from fibers to fabrics in each wavelength. Our approach requires only five pairs of monochromatic fibers and their corresponding fabrics for training, significantly reducing the complexity and resources needed. This color transfer model integrates our previously developed spectral calibration model for both colored and noncolored fibers based on the single-constant Kubelka–Munk (KM) law to achieve direct and accurate color prediction for colored spun fabrics. Experimentally, a total of 87 cotton colored spun fabrics were prepared to test this new method. Satisfactorily, remarkable average and maximum color differences of merely 0.78 and 1.82 CIEDE2000 units were obtained, which were significantly better than that of the single-constant KM model (∼6.01, 12.27), two-constant KM model (∼1.17, 8.94), Stearns–Noechel model (∼1.02, 4.09), and Friele model (∼1.43, 2.87). These results demonstrate that the proposed color transfer model significantly improves accuracy and efficiency in color prediction for colored spun fabrics, offering a more reliable alternative to traditional models.