Fast and accurate measurement of colors in multicolored prints using commercial instruments or existing computer vision systems remains a challenge due to limitations in image segmentation methods and the size and complexity of the colored patterns. To determine the colorimetric attributes (L*a*b*) of multicolored materials, an approach based on global color correction and an effective unsupervised image segmentation is presented. The colorimetric attributes of all patches in a ColorChecker chart were measured spectrophotometrically, and an image of the chart was also captured. Images were segmented using a modified Chan-Vese method, and the sRGB values of each patch were extracted and then transformed into L*a*b* values. In order to optimize the transformation process, the performance of 10 models was examined by minimizing the average color differences between measured and calculated colorimetric values. To assess the performance of the model, a set of printed samples was employed and the color differences between the predicted and measured L*a*b* values of samples were compared. The results show that the modified Chan-Vese method, with suitable settings, generates satisfactory segmentation of the printed images with mean and maximum ΔE00 values of 2.43 and 4.28 between measured and calculated values.