INTERNATIONAL ORGANIZATION FOR S

Source:	Jörn Ostermann, Touradj Ebrahimi
Title:	Results of the First Workshop on Future Directions in Video Compression
Status:	*Approved*

During its 72^nd meeting MPEG held its first ‘Workshop on Future Directions in Video Compression’ on April 21^st, 2005 in Busan Korea. The workshop was open to the interested public and attracted about 250 participants. The program of the workshop contained 7 talks covering all aspects of a successful deployment of future video systems and applications. The program of the workshop and the slides of the presentations are part of this document.

During the workshop a few trends to be considered in future developments were identified:

In the next few years, the use of high resolution cameras and displays will be seen in the consumer market. HDTV with a resolution of about 1020 lines and 1980 pels/line will be competing with new formats providing higher resolutions and better color rendition. Multi-view formats will become increasingly available.

Advancement in video coding based on today’s video coding standards will be driven by better motion compensation, texture synthesis complementing texture coding using DCT and native RGB compression to allow for better colour rendition.

Distributed source coding was identified as a new compression scheme suitable for applications that require low-cost video encoders as well as a scheme for providing robustness in video coding. This technology is still in its infancy and will require at least 5 years to reach a mature state. X-lets like edge-lets, ridge-lets, and others, are an extension of wavelets that might provide a performance competitive with today’s video coding standards.

Most experts present at the workshop expect future mobile networks to be designed such that their behavior towards the application is not very different to that of wired networks. Bit errors will be concealed by the network. The packet loss rates are about 0.1% with occasional break down of the connection. Some experts believe that it will still be possible to receive packets with bit errors given that some protocols as defined in 3GPP and the IETF Datagram Congestion Control Protocol and IETF UDP-lite specification.

The dominant protocol on mobile networks will be based on IP. Available data rates are increasing fast in the next ten years.

Implementation cost of a video coding standard is an important factor to be considered when developing the standard. Tools are becoming available that allow for an automatic implementation complexity based on C-code.

The Second Workshop on Future Directions in Video Compression will be held on October 16^th 2005 in Nice

Programme of first Workshop

Dr. Frank Bossen, NTT DoCoMo Labs USA	Next generation mobile networks and their implications for video applications
Kannan Ramchandran, Rohit Puri, Abhik Majumdar, Jiajun Wang, Marco Tagliasacchi+,* University of California, Berkeley, + Politecnico di Milano	Video coding based on distributed source coding principles
Masayuki TANIMOTO, Nagoya University, Japan	Ray-Based Image System
Woo-Shik Kim†, Dae-Sung Cho†, Hyun Mun Kim†, and Shijun Sun‡, †Computing Lab, Samsung Advanced Institute of Technology, ‡Sharp Labs of America	New High Fidelity RGB Video Coding
Joern Ostermann, University of Hannover	Hybrid Coding: Where can future gains come from?
Marco Mattavelli(1), Robert Turney(2), (1)Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, (2)Xilinx Research Labs	Which Metrics for Developing Low-Complexity Video Compression Solutions
Touradj Ebrahimi, EPFL	Back to the Future of Video Compression