CONTENT-AWARE SUPERVISION FOR DIFFUSION-BASED RESTORATION OF EXTREMELY COMPRESSED BACKGROUND FOR VCM

Le Thi Hue Dao, An Gia Vien, Jooyoung Lee, Seyoon Jeong, Naeun Yang, Chul Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We propose content-aware supervision (CAS) techniques for diffusion-based restoration of an extremely compressed background for video coding for machines (VCM). First, we develop a CAS block to exploit prior information in an input image to reconstruct the noisy image, which is used as the input for the pretrained diffusion model. Then, we construct a refinement block to guide the pretrained diffusion model at each diffusion step by incorporating a degradation model and correction gradient estimation. Experimental results demonstrate the proposed algorithm outperforms state-of-the-art algorithms.

Original languageEnglish
Title of host publication2024 IEEE International Conference on Image Processing, ICIP 2024 - Proceedings
PublisherIEEE Computer Society
Pages1683-1689
Number of pages7
ISBN (Electronic)9798350349399
DOIs
StatePublished - 2024
Event31st IEEE International Conference on Image Processing, ICIP 2024 - Abu Dhabi, United Arab Emirates
Duration: 27 Oct 202430 Oct 2024

Publication series

NameProceedings - International Conference on Image Processing, ICIP
ISSN (Print)1522-4880

Conference

Conference31st IEEE International Conference on Image Processing, ICIP 2024
Country/TerritoryUnited Arab Emirates
CityAbu Dhabi
Period27/10/2430/10/24

Keywords

  • diffusion model
  • Image generation
  • image restoration
  • video coding for machines (VCM)

Fingerprint

Dive into the research topics of 'CONTENT-AWARE SUPERVISION FOR DIFFUSION-BASED RESTORATION OF EXTREMELY COMPRESSED BACKGROUND FOR VCM'. Together they form a unique fingerprint.

Cite this