Display management for high dynamic range video

A method for the display management of images with one or more computer processors, the method comprising:

• accessing an input image in a first color space with a first enhanced dynamic range (EDR) of a source display device;
• performing a color transformation operation on the input image to generate a first output image (112) in a perceptually-corrected IPT color space (IPT-PQ), wherein the color transformation operation transforming the input image from the first color space to the IPT-PQ space is based at least in part on applying non-linear perceptual quantization (PQ) to the input image;
• performing a non-linear tone-mapping operation on an intensity (I) component of the first output image to generate an intensity component of a first tone-mapped image with a second dynamic range of a target display device, the second dynamic range of the target display device being different than the first dynamic range of the source display device;
• performing a detail preservation operation to generate an intensity component of a second output image in response to the intensity components of the first output image and the first tone-mapped image;
• performing a saturation adjustment operation on color (P, T) components of the second output image to generate color components of a second tone-mapped image, wherein the saturation adjustment operation is determined by the intensity component of the first output image and the intensity component of either the second output image or the first tone-mapped image;
• and
• performing a color gamut mapping operation on the second tone-mapped image to generate a third output image;
• causing a display mapped image generated from the third output image to be rendered on the target display device;
• wherein performing the detail preservation operation further comprises computing
• IS=Io−F((Io−Im),H),
• where F(I,H) denotes applying to an image I a filter with kernel H, Io denotes intensity pixel values of the first output image, Im denotes intensity pixel values of the first tone-mapped output image, and Is denotes intensity pixel values of the second output image.