PROCEEDINGS OF THE ITE ANNUAL CONVENTION
Online ISSN : 2424-2292
Print ISSN : 1343-1846
ISSN-L : 1343-1846
Proceedings of the ITE Annual Convention 2018
Conference information

A Study on Block-based Image Coding Applying Completion by Generative Adversarial Networks
*Shota ORIHASHIShinobu KUDOMasaki KITAHARAAtsushi SHIMIZU
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Pages 31B-1-

Details
Abstract
In this paper, we propose a block-based image coding technique applying completion which is constructed by framework of generative adversarial networks. Our method is realized by pre-processing and post-processing of HEVC. While the method is processed in coding tree units, region of image completion can be specified in smaller blocks.
Content from these authors
© 2018 The Institute of Image Information and Television Engineers
Previous article Next article
feedback
Top