Skip to main navigation Skip to search Skip to main content

Region of interest signalling for encrypted JPEG images

  • Department of Computer Science, University of Salzburg

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

Abstract

We propose and evaluate different methods to signal position and size of encrypted RoIs (Regions of Interest) in JPEG images. After discussing various design choices regarding the encoding of RoI coordinates with a minimal amount of bits, we discuss both, existing and newly proposed approaches to signal the encoded coordinates inside JPEG images. By evaluating the different signalling methods on various data sets, we show that several of our proposed encoding methods outperform JBIG in this special use case. Furthermore, we show that one of our proposed signalling methods allows length-preserving lossless signalling, i.e., storing RoI coordinates in a format-compliant way inside the JPEG images without quality loss or change of file size. © 2013 ACM.
Original languageEnglish
Title of host publicationIH&MMSec '13: Proceedings of the first ACM workshop on Information hiding and multimedia security
PublisherAssociation for Computing Machinery
Pages165-174
Number of pages10
ISBN (Print)978-1-4503-2081-8
DOIs
Publication statusPublished - 2013
Event1st ACM Workshop on Information Hiding and Multimedia Security, IH and MMSec 2013 - Montpellier, France
Duration: 17 Jun 201319 Jun 2013

Workshop

Workshop1st ACM Workshop on Information Hiding and Multimedia Security, IH and MMSec 2013
Abbreviated titleIH and MMSec 2013
Country/TerritoryFrance
CityMontpellier
Period17/06/1319/06/13

Keywords

  • coordinates
  • encoding
  • JPEG
  • region of interest
  • signalling
  • Encoding methods
  • Format-compliant
  • JPEG image
  • Quality loss
  • Region of interest
  • Regions of interest
  • Cryptography
  • Encoding (symbols)
  • Signaling
  • Image coding

Fingerprint

Dive into the research topics of 'Region of interest signalling for encrypted JPEG images'. Together they form a unique fingerprint.

Cite this