Piecewise Cross Color Channel Predictor

A method comprising:

• accessing a first image and a second image, each of the images comprising one or more color channels, each of the images comprising a plurality of pixels, each pixel having a respective pixel value for each of the one or more color channels, wherein the second image has a dynamic range that is higher than a dynamic range of the first image;
• segmenting at least one color channel of the first image into two or more non-overlapping color channel segments using a set of boundary points, wherein each color channel segment corresponds to two consecutive boundary points, and wherein the pixel values of the color channel which are between two consecutive boundary points are assigned to the corresponding color channel segment;
• and
• for a color channel segment of the first image:
• selecting a piece-wise cross-color channel (PCCC) prediction model for the color channel segment from one or more PCCC prediction models wherein a predicted pixel value of a pixel of the second image in one color channel is expressed as a combination of at least the respective pixel values for all color channels of the pixel within the first image having the same pixel coordinates as the pixel of the second image;
• solving for prediction parameters of the selected prediction model;
• computing an output color channel segment based on the first image, the second image, and the prediction parameters of the selected prediction model;
• and
• outputting the prediction parameters of the selected prediction model for use by a decoder.