eduzhai > Applied Sciences > Engineering >

High-Fidelity Generative Image Compression

  • Save

... pages left unread,continue reading

Document pages: 20 pages

Abstract: We extensively study how to combine Generative Adversarial Networks andlearned compression to obtain a state-of-the-art generative lossy compressionsystem. In particular, we investigate normalization layers, generator anddiscriminator architectures, training strategies, as well as perceptual losses.In contrast to previous work, i) we obtain visually pleasing reconstructionsthat are perceptually similar to the input, ii) we operate in a broad range ofbitrates, and iii) our approach can be applied to high-resolution images. Webridge the gap between rate-distortion-perception theory and practice byevaluating our approach both quantitatively with various perceptual metrics,and with a user study. The study shows that our method is preferred to previousapproaches even if they use more than 2x the bitrate.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×