The Self-Adaptive Local Nonlinear Regression Method for Color Reproduction

Look-up Table (LUT) is one of the popular methods for color reproduction because of its high precision. With the aim to improve the performance of color reproduction, the self-adaptive local nonlinear regression method is proposed in the paper. On the base of self-adaptive local linear regression, the new method has introduced Maximum Entropy Estimation and nonlinear regression to establish a high-precision three dimensional LUT. The result of experiment shows that the self-adaptive local nonlinear regression method works better than the self-adaptive local linear regression method. The image corrected by the new method appears smoothly, has natural transition between colors and distinct details.