Unsupervised training of neural network for high dynamic range image compression

At least one non-transitory computer readable storage medium having instructions encoded thereon that, when executed by one or more processors, cause a process to be carried out for training a neural network to perform high dynamic range (HDR) image compression, the process comprising:

• performing global tone mapping on an HDR training image to generate a low dynamic range (LDR) training image;
• applying the neural network to the HDR training image and the LDR training image, the neural network to generate a delta image, the delta image representing image detail lost in the global tone mapping operation;
• summing the delta image with the LDR training image to generate an output training image;
• calculating a contrast loss, the contrast loss based on the output training image and the HDR training image;
• calculating a compression loss, the compression loss based on the output training image and the LDR training image;
• and
• calculating a loss function based on from a weighted sum of the contrast loss and the compression loss.

A system for perform high dynamic range (HDR) image compression, the system comprising:

• a global tone mapping circuit to transform an HDR image into a low dynamic range (LDR) image;
• a neural network trained to generate a delta image based on the HDR image and the LDR image, the delta image representing image detail lost in the global tone mapping operation, wherein the training of the neural network is an unsupervised training process to minimize a loss function calculated from a weighted sum of a contrast loss and a compression loss;
• and
• a summing circuit to sum the LDR image and the delta image to generate a compressed output image.