Multi-Realism Image Compression With a Conditional Generator

A method, comprising:

• receiving a source image;
• encoding the source image using an autoencoder to generate an encoded representation, wherein the autoencoder comprises a hyperprior-based architecture;
• conditioning a generator on a realism factor to produce a reconstructed image from the encoded representation, wherein the realism factor is adjustable to control a level of synthesized content in the reconstructed image;
• and
• storing or transmitting the encoded representation.

A non-transitory computer-readable storage medium, comprising executable instructions that, when executed by a processor, cause performance of operations that perform a method comprising:

• obtaining a bitstream that includes an encoded representation of a source image;
• receiving a realism factor indicative of an amount of synthesized content in a reconstructed image of the source image;
• inputting the realism factor and the encoded representation to a decoder to obtain the reconstructed image of the source image;
• and
• storing or displaying the reconstructed image.