Published June 11, 2023 | Version v1
Conference paper Open

Towards an international standard to establish trust in media production, distribution and consumption

  • 1. Vrije Universiteit Brussel and imec
  • 2. The Australian National University
  • 3. Centre for Research and Technology Hellas
  • 4. Instituto de Telecomunicações
  • 5. University College London

Description

Advances in media content manipulation and artificially generated content pose new challenges to the assessment of media authenticity. While automated detection methods can provide meaningful insights and decision support in some scenarios, they cannot provide trustworthy and comprehensive information about the origin and provenance of media assets. Therefore, a longer-term approach should rather focus on secure and interoperable annotations related to the creation and provenance of media. In October 2020, the JPEG Committee initiated a standardization exploration named “JPEG Fake Media” to address these needs. Subsequently, since many of the requirements, for example related to secure annotation and identification of media assets, are also relevant to achieve interoperability in Non-Fungible Tokens (NFTs) an additional exploration was initiated, specifically focused on standardization needs for NFTs. In April 2022 a first Call for Proposals on JPEG Fake Media was issued. Based on the responses to the call, a new standardization project named JPEG Trust was initiated to specify an interoperable framework for establishing trust in media production, distribution, and consumption. This paper presents the journey of JPEG to leverage formal methods of standardization in this context, starting from the initial JPEG Fake Media exploration, followed by the subsequent consideration of NFT use cases and requirements, through to the commencement of the new JPEG Trust international standard.

Files

DSP2023-55.pdf

Files (106.6 kB)

Name Size Download all
md5:4bc647ab49a7c24ca5a0d05192135727
106.6 kB Preview Download

Additional details

Funding

vera.ai – vera.ai: VERification Assisted by Artificial Intelligence 101070093
European Commission