Research on color prediction model based on radial basis function neural network for 3D printing

Abstract

3D printing technology has enabled full-color printing. Research on factors and controls affecting color reproduction in full-color 3D printing has attracted more and more attention. Among them, the background factor and the finish type are two key factors which affect the color reproduction of 3D printing. In this paper, the Poly jet J750 was selected as the output device and one color prediction model based on RBF (Radial Basis Function) neural network was proposed to predict color reproduction for color 3D printing. It contains three input variables, such as the original chromatic coordinates of the digital files, i.e., CIE1976L*a*b*, the finish type of the printed color blocks, and the background factor in the proposed model. The measured color chromatic coordinates of the printed color blocks were set as output variables for the model. The prediction performance of the model was evaluated by using the mean absolute error (MAE), the coefficient of determination (R 2 ) and the CIEDE2000 color difference. It shows that the proposed model had a good prediction effect. In addition, it also shows that the white background has a better effect on the color prediction than those of the black and neutral gray background. And for the finish type, the Glossy has a better effect on the color prediction than that of the Matte.