Title :
Rate-distortion optimizations for region and object based wavelet video coding
Author :
Yang, Yan ; Hemami, Sheila S.
Author_Institution :
Aware Inc., Bedford, MA, USA
fDate :
Oct. 29 2000-Nov. 1 2000
Abstract :
This paper provides rate-distortion optimizations for both region- and object-based video encoding with a motion-compensated wavelet coder. The rate-distortion optimization which jointly selects all coding parameters for a frame is first applied to single video object plane (SVOP) coding via the Lagrange multiplier method. The optimal bit allocation for multiple regions of interest (MROI) coding is then achieved using a weighted Lagrange multiplier technique. Finally, the generalized R-D optimization for multiple video object plane (MVOP) coding is realized with the weighted technique by including special handling of object boundary blocks, which provides the additional capability of object scalability. The weights allow direct selection of relative distortions in the different regions or object planes, and simulations on MPEG-4 test sequences demonstrate the desired inhomogeneity in reconstruction quality.
Keywords :
image sequences; motion compensation; motion estimation; optimisation; rate distortion theory; transform coding; video coding; wavelet transforms; Lagrange multiplier method; MPEG-4 test sequences; MROI coding; MVOP coding; SVOP coding; generalized R-D optimization; motion-compensated wavelet coder; multiple regions of interest coding; multiple video object plane coding; object boundary block; object scalability; object-based video coding; optimal bit allocation; rate-distortion optimization; reconstruction quality; region-based video coding; single video object plane coding; wavelet video coding; weighted technique; Bit rate; Constraint optimization; Design for disassembly; Encoding; Lagrangian functions; Motion estimation; Optimization methods; Rate-distortion; Scalability; Video coding;
Conference_Titel :
Signals, Systems and Computers, 2000. Conference Record of the Thirty-Fourth Asilomar Conference on
Conference_Location :
Pacific Grove, CA, USA
Print_ISBN :
0-7803-6514-3
DOI :
10.1109/ACSSC.2000.911214