Share Email Print
cover

Proceedings Paper

Error-resilient low-delay H.264/802.11 transmission via cross-layer coding with feedback channel
Author(s): Tuan-Kiang Chiew; Paul Hill; Pierre Ferre; Dimitris Agrafiotis; James T. H. Chung-How; Andy Nix; David R. Bull
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

We propose a method of providing error resilient H.264 video over 802.11 wireless channels by using a feedback mechanism which does not incur an additional delay typically found in ARQ-type feedback. Our system uses the TCP/IP and UDP/IP protocols, located between the medium access control (MAC) layer of 802.11, and the H.264 video application layer. The UDP protocol is used to transfer time sensitive video data without delay; however, packet losses introduce excessive artifacts which propagate to subsequent frames. Error resilience is achieved by a feedback mechanism-the decoder conveys the packet-loss information as small TCP packets to the video source as negative acknowledgements. By using multiple reference frames, slice-based coding and timely intra-refresh, the encoder makes use of this feedback information to perform subsequent temporal prediction without propagating the error to future frames. We take static measurements of the actual channel and use the packet loss and delay patterns to test our algorithms. Simulations show an improvement of 0.5~5 dB in PSNR over plain UDP-based video transmission. Our method improves the overall quality of service of interactive video transmission over wireless LAN; it can be used as a model for future media-aware wireless network protocol designs.

Paper Details

Date Published: 24 June 2005
PDF: 12 pages
Proc. SPIE 5960, Visual Communications and Image Processing 2005, 59605H (24 June 2005); doi: 10.1117/12.633376
Show Author Affiliations
Tuan-Kiang Chiew, Institute for Infocomm Research (Singapore)
Paul Hill, Univ. of Bristol (United Kingdom)
Pierre Ferre, Univ. of Bristol (United Kingdom)
Dimitris Agrafiotis, Univ. of Bristol (United Kingdom)
James T. H. Chung-How, ProVision Communication Technologies Ltd. (United Kingdom)
Andy Nix, Univ. of Bristol (United Kingdom)
ProVision Communication Technologies Ltd. (United Kingdom)
David R. Bull, Univ. of Bristol (United Kingdom)
ProVision Communication Technologies Ltd. (United Kingdom)


Published in SPIE Proceedings Vol. 5960:
Visual Communications and Image Processing 2005
Shipeng Li; Fernando Pereira; Heung-Yeung Shum; Andrew G. Tescher, Editor(s)

© SPIE. Terms of Use
Back to Top