loading...
Bit Allocation Based on Motion Vector Analysis for H.264/AVC
Snowbird, Utah March 27-March 29
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/DCC.2007.172007 Data Compression Conference (DCC ...
 This Article 
 
PDF
HTML
 
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Hussain M. Mohammed, Fraunhofer Institute for Integrated Circuits, Germany
Nikolaus Farber, Fraunhofer Institute for Integrated Circuits, Germany
Herbert Thoma, Fraunhofer Institute for Integrated Circuits, Germany
In this paper we present a bit allocation approach based on motion vector analysis for improved Rate-Distortion performance. Bit allocation is done at the macroblock level such that the macroblock with high priority is coded finely and the one with low priority is coded coarsely. In order to calculate macroblock priorities, first reference counts for each pixel are determined through motion vector analysis. A reference count of a pixel is defined as total number of pixels in the remaining GOP which use that pixel as a reference. Then macroblock wise reference counts is obtained by summing the pixel wise reference counts, which are then scaled and prioritized. Based upon the priority value the given fixed quant of each frame is modulated at the macroblock level. The algorithm is applied to H.264/AVC encoding and PSNR gains of up to 1.4 dB are achieved.
Citation:
Hussain M. Mohammed, Nikolaus Farber, Herbert Thoma, "Bit Allocation Based on Motion Vector Analysis for H.264/AVC," dcc, pp.313-322, 2007 Data Compression Conference (DCC'07), 2007
Usage of this product signifies your acceptance of the Terms of Use.