TY - GEN
T1 - A hierarchical control scheme for energy quota distribution in Hybrid Distributed Video Coding
AU - Khan, Muhammad Usman Karim
AU - Shafique, Muhammad
AU - Henkel, Jörg
N1 - Copyright:
Copyright 2012 Elsevier B.V., All rights reserved.
PY - 2012
Y1 - 2012
N2 - Hybrid Distributed Video Coding (HDVC) has emerged as an attractive solution for energy and resource constrained video processing systems in different domains like distributed video sensors, mobile devices with dynamically varying energy levels, or hybrid paradigms where mobile devices are communicating with high-end servers. In this paper, we present a hierarchical control scheme for determining and distributing the energy quota in HDVC encoder and decoder under scenarios of dynamically varying energy levels and user constraints. The energy quotas are computed and con-trolled online at various hierarchical levels (like group of frames, frames, and macroblocks), while jointly accounting for the computation and transmission energy. For energy reduction, the processing of selective regions is intelligently distributed at both encoder and decoder sides, considering the texture and motion properties of different video regions and available spatial/temporal correlations. Experimental results demonstrate that our scheme provides on average 20%-25% reduced energy consumption com-pared to state-of-the-art HDVC schemes. The key to high energy efficiency is to leverage the video content properties.
AB - Hybrid Distributed Video Coding (HDVC) has emerged as an attractive solution for energy and resource constrained video processing systems in different domains like distributed video sensors, mobile devices with dynamically varying energy levels, or hybrid paradigms where mobile devices are communicating with high-end servers. In this paper, we present a hierarchical control scheme for determining and distributing the energy quota in HDVC encoder and decoder under scenarios of dynamically varying energy levels and user constraints. The energy quotas are computed and con-trolled online at various hierarchical levels (like group of frames, frames, and macroblocks), while jointly accounting for the computation and transmission energy. For energy reduction, the processing of selective regions is intelligently distributed at both encoder and decoder sides, considering the texture and motion properties of different video regions and available spatial/temporal correlations. Experimental results demonstrate that our scheme provides on average 20%-25% reduced energy consumption com-pared to state-of-the-art HDVC schemes. The key to high energy efficiency is to leverage the video content properties.
KW - Adaptivity
KW - Distributed Video Coding (DVC)
KW - Energy budgeting
KW - Hybrid Distributed Video Coding (HDVC)
KW - Low-power
KW - Motion estimation
KW - Video coding
UR - http://www.scopus.com/inward/record.url?scp=84869054257&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84869054257&partnerID=8YFLogxK
U2 - 10.1145/2380445.2380519
DO - 10.1145/2380445.2380519
M3 - Conference contribution
AN - SCOPUS:84869054257
SN - 9781450314268
T3 - CODES+ISSS'12 - Proceedings of the 10th ACM International Conference on Hardware/Software-Codesign and System Synthesis, Co-located with ESWEEK
SP - 483
EP - 492
BT - CODES+ISSS'12 - Proceedings of the 10th ACM International Conference on Hardware/Software-Codesign and System Synthesis, Co-located with ESWEEK
T2 - 10th ACM International Conference on Hardware/Software-Codesign and System Synthesis, CODES+ISSS 2012, Co-located with 8th Embedded Systems Week, ESWEEK 2012
Y2 - 7 October 2012 through 12 October 2012
ER -