Reconstruction Distortion of Learned Image Compression with Imperceptible Perturbations

Yang SUI, Zhuohang LI, Ding DING, Xiang PAN, Xiaozhong XU, Shan LIU, Zhenzhong CHEN*

*Corresponding author for this work

Research output: Book Chapters | Papers in Conference ProceedingsConference paper (refereed)Researchpeer-review

1 Citation (Scopus)

Abstract

In this paper, we introduce an imperceptible adversarial attack approach designed to effectively degrade the reconstruction quality of LIC, resulting in the reconstructed image being severely disrupted by noise where identifying any object in the reconstructed image is virtually impossible. More specifically, we generate adversarial examples by introducing a Frobenius norm-based loss function to maximize the discrepancy between original images and reconstructed images from adversarial examples in order to corrupt the reconstructed image severely.
Original languageEnglish
Title of host publication2024 Data Compression Conference, DCC 2024: Proceedings
EditorsAli BILGIN, James E. FOWLER, Joan SERRA-SAGRISTA, Yan YE, James A. STORER
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages583
Number of pages1
ISBN (Electronic)9798350385878
DOIs
Publication statusPublished - 2024
Externally publishedYes
Event2024 Data Compression Conference - Snowbird, United States
Duration: 19 Mar 202422 Mar 2024

Publication series

NameData Compression Conference: Proceedings
PublisherIEEE
ISSN (Print)1068-0314
ISSN (Electronic)2375-0359

Conference

Conference2024 Data Compression Conference
Abbreviated titleDCC 2024
Country/TerritoryUnited States
CitySnowbird
Period19/03/2422/03/24

Bibliographical note

Publisher Copyright:
© 2024 IEEE.

Keywords

  • adversarial attack
  • image compression
  • invisible adversarial attack
  • learned image compression
  • robustness

Fingerprint

Dive into the research topics of 'Reconstruction Distortion of Learned Image Compression with Imperceptible Perturbations'. Together they form a unique fingerprint.

Cite this