Optimizing distributed source coding for interactive multiview video streaming over lossy networks

Zhi Liu, Gene Cheung, Yusheng Ji

Research output: Contribution to journalArticle

39 Citations (Scopus)

Abstract

In interactive multiview video streaming (IMVS), a user observes one view at a time, but can periodically switch to a desired neighboring captured view as the video is played back in time. Previous IMVS works focus on efficient compression techniques that facilitate interactive view switching. In this paper, in addition to the loss-resilient aspect during network streaming we address how to design efficient coding tools and optimize frame structure for transmission to facilitate view switching and contain error propagation in differentially coded video due to packet losses. We first design a new unified distributed source coding (uDSC) frame - a new coding tool that simultaneously offers view switching and loss-resilient capabilities - for periodic insertion into the multiview frame structure. After inserting uDSC-frames into the coding structure, we schedule packets for network transmission in a rate-distortion optimal manner for both wireless multicast and wired unicast streaming scenarios. For wireless multicast over a Gilbert-Elliott loss model, frames in a group of pictures are packetized and reordered, so that uDSC frames are correctly decoded with high probability, mitigating error propagation. For wired unicast, we use a Markov decision process to optimize packet transmission to minimize expected distortion given a bandwidth constraint. Experimental results show that systems that insert uDSC frames and optimize packet transmission can outperform other competing coding schemes by up to 2.8 and 11.6 dB in wireless multicast and wired unicast streaming scenarios, respectively.

Original languageEnglish
Article number6542004
Pages (from-to)1781-1794
Number of pages14
JournalIEEE Transactions on Circuits and Systems for Video Technology
Volume23
Issue number10
DOIs
Publication statusPublished - 2013
Externally publishedYes

Fingerprint

Video streaming
Electric power transmission networks
Packet loss
Switches
Bandwidth

Keywords

  • Distributed source coding
  • interactive multiview video streaming
  • Markov decision process

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Media Technology

Cite this

Optimizing distributed source coding for interactive multiview video streaming over lossy networks. / Liu, Zhi; Cheung, Gene; Ji, Yusheng.

In: IEEE Transactions on Circuits and Systems for Video Technology, Vol. 23, No. 10, 6542004, 2013, p. 1781-1794.

Research output: Contribution to journalArticle

@article{9ddf5eebf4474bdebcba16f7134a9697,
title = "Optimizing distributed source coding for interactive multiview video streaming over lossy networks",
abstract = "In interactive multiview video streaming (IMVS), a user observes one view at a time, but can periodically switch to a desired neighboring captured view as the video is played back in time. Previous IMVS works focus on efficient compression techniques that facilitate interactive view switching. In this paper, in addition to the loss-resilient aspect during network streaming we address how to design efficient coding tools and optimize frame structure for transmission to facilitate view switching and contain error propagation in differentially coded video due to packet losses. We first design a new unified distributed source coding (uDSC) frame - a new coding tool that simultaneously offers view switching and loss-resilient capabilities - for periodic insertion into the multiview frame structure. After inserting uDSC-frames into the coding structure, we schedule packets for network transmission in a rate-distortion optimal manner for both wireless multicast and wired unicast streaming scenarios. For wireless multicast over a Gilbert-Elliott loss model, frames in a group of pictures are packetized and reordered, so that uDSC frames are correctly decoded with high probability, mitigating error propagation. For wired unicast, we use a Markov decision process to optimize packet transmission to minimize expected distortion given a bandwidth constraint. Experimental results show that systems that insert uDSC frames and optimize packet transmission can outperform other competing coding schemes by up to 2.8 and 11.6 dB in wireless multicast and wired unicast streaming scenarios, respectively.",
keywords = "Distributed source coding, interactive multiview video streaming, Markov decision process",
author = "Zhi Liu and Gene Cheung and Yusheng Ji",
year = "2013",
doi = "10.1109/TCSVT.2013.2269019",
language = "English",
volume = "23",
pages = "1781--1794",
journal = "IEEE Transactions on Circuits and Systems for Video Technology",
issn = "1051-8215",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "10",

}

TY - JOUR

T1 - Optimizing distributed source coding for interactive multiview video streaming over lossy networks

AU - Liu, Zhi

AU - Cheung, Gene

AU - Ji, Yusheng

PY - 2013

Y1 - 2013

N2 - In interactive multiview video streaming (IMVS), a user observes one view at a time, but can periodically switch to a desired neighboring captured view as the video is played back in time. Previous IMVS works focus on efficient compression techniques that facilitate interactive view switching. In this paper, in addition to the loss-resilient aspect during network streaming we address how to design efficient coding tools and optimize frame structure for transmission to facilitate view switching and contain error propagation in differentially coded video due to packet losses. We first design a new unified distributed source coding (uDSC) frame - a new coding tool that simultaneously offers view switching and loss-resilient capabilities - for periodic insertion into the multiview frame structure. After inserting uDSC-frames into the coding structure, we schedule packets for network transmission in a rate-distortion optimal manner for both wireless multicast and wired unicast streaming scenarios. For wireless multicast over a Gilbert-Elliott loss model, frames in a group of pictures are packetized and reordered, so that uDSC frames are correctly decoded with high probability, mitigating error propagation. For wired unicast, we use a Markov decision process to optimize packet transmission to minimize expected distortion given a bandwidth constraint. Experimental results show that systems that insert uDSC frames and optimize packet transmission can outperform other competing coding schemes by up to 2.8 and 11.6 dB in wireless multicast and wired unicast streaming scenarios, respectively.

AB - In interactive multiview video streaming (IMVS), a user observes one view at a time, but can periodically switch to a desired neighboring captured view as the video is played back in time. Previous IMVS works focus on efficient compression techniques that facilitate interactive view switching. In this paper, in addition to the loss-resilient aspect during network streaming we address how to design efficient coding tools and optimize frame structure for transmission to facilitate view switching and contain error propagation in differentially coded video due to packet losses. We first design a new unified distributed source coding (uDSC) frame - a new coding tool that simultaneously offers view switching and loss-resilient capabilities - for periodic insertion into the multiview frame structure. After inserting uDSC-frames into the coding structure, we schedule packets for network transmission in a rate-distortion optimal manner for both wireless multicast and wired unicast streaming scenarios. For wireless multicast over a Gilbert-Elliott loss model, frames in a group of pictures are packetized and reordered, so that uDSC frames are correctly decoded with high probability, mitigating error propagation. For wired unicast, we use a Markov decision process to optimize packet transmission to minimize expected distortion given a bandwidth constraint. Experimental results show that systems that insert uDSC frames and optimize packet transmission can outperform other competing coding schemes by up to 2.8 and 11.6 dB in wireless multicast and wired unicast streaming scenarios, respectively.

KW - Distributed source coding

KW - interactive multiview video streaming

KW - Markov decision process

UR - http://www.scopus.com/inward/record.url?scp=84885617275&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84885617275&partnerID=8YFLogxK

U2 - 10.1109/TCSVT.2013.2269019

DO - 10.1109/TCSVT.2013.2269019

M3 - Article

AN - SCOPUS:84885617275

VL - 23

SP - 1781

EP - 1794

JO - IEEE Transactions on Circuits and Systems for Video Technology

JF - IEEE Transactions on Circuits and Systems for Video Technology

SN - 1051-8215

IS - 10

M1 - 6542004

ER -