An Efficient Color Grading Method Using Warped Color Space Mapping Based on Color Distribution Deformation

As a novel approach to color correction based on tone-mapping methods, we propose a method that distorts the color space using lattice deformation, which deforms the shape of the color distribution of the input image, and changes the colors of the output image. In conventional tone-mapping methods for color images, the bottleneck is the process of moving color positions along each arbitrary axis set in the color space. Each time an axis is selected, by using all pixel values, the tone-mapping function needs to be computed. Additionally, the colors of all pixel values need to be updated regularly. In the proposed method, the same processing is performed by mapping similar pixel values (RGB values) to the representative color position (RGB coordinates) in the RGB color space. With this approach, it may be possible to reduce the number of computations using all pixel values. To demonstrate the effectiveness of our approach with conventional methods, we show the results of an improved version of a color grading method based on the deformation of color distributions.