DIOMEDES: Distribution of Multi-view Entertainment Using Content Aware Delivery Systems

DIOMEDES was a collaborative project, funded under the European Union FP7 ICT Networked Media research programme, started in February 2010 and ended in July 2012. It integrated 7 partners from 5 countries across Europe and was led by the I-Lab team (the current members of the IDT at Loughborough University London).

The total funding of the project was €3.6M. The project focused on investigating and developing new methods for the compression and delivery of multi-view video and multi-channel audio to users.

The DIOMEDES approach was to develop a 3D Peer-to-Peer (P2P) distribution system, which was designed jointly with novel scalable video compression techniques. The compression, distribution, and security technologies developed within DIOMEDES were demonstrated in the form of an integrated test-bed. This test-bed included a user terminal device prototype, consisting of a cluster of PCs to run all prototype software modules including P2P streaming client in combination with an adaptation decision engine, audio-visual synchronisation buffer, authentication client, and the audio and video media players that rendered the scalable coded multi-view content on a variety of 3D displays.

DIOMEDES also comprised the development of a real-time interactive free-viewpoint 3D viewing experience by novel, efficient multi-view plus depth video renderer.

The project’s major features comprised:

  • 3D Quality of Experience (QoE) and 3D audio‐visual attention modelling – It was ensured that human perception of 3D multimedia was considered during the optimisation of 3D media delivery services. This work included optimised adaptation to network congestion and for displays/TVs with different capabilities. For this purpose, novel Key Performance Index (KPI) computed at the server side were delivered to users in the video transport streams to improve adaptation process.
  • Multi-view HD video and spatial audio capture and post-processing, including depth extraction and content-aware pre-processing for better compression performance.
  • Visual attention based scalable 3D video coding and content, and network-aware P2P distribution – the designed P2P 3D media distribution system coupled with the novel scalable codec is applicable to the current internet infrastructure, supported by most Internet Service Providers (ISPs). Distribution of media packets across peers was augmented to reduce the loads on servers.
  • Content and context-aware 3D multi-view video adaptation decision taking mechanisms incorporating various parameters.
  • Joint use of terrestrial broadcast and IP networks for synchronous 3D media delivery – multi‐view content was distributed to the user through multiple paths to balance the information load as well as to ensure a more robust 3D multimedia delivery at homes.
  • “Controlled” sharing of content (including user‐generated content), i.e. ensuring only the authorised content was distributed, as well as content authentication and access control.