Generating optimal look-up tables to achieve complex color space transformations

Color space transformations are very common in digital imaging and display systems. Their complex mathematics requires hardware friendly implementations to make them practical. In this paper, look-up tables (LUT) are described for implementation of the color transformations. We show that this is possible because an analytical model of the transform is available for training and cross-validation. We identify parameters that affect the performance of a LUT-based system and then formulate the solution as an optimization problem.