DocumentCode :
318221
Title :
Reduced-complexity rate-distortion optimization of multiresolution motion field and prediction residual
Author :
Subramanian, Pattabiraman ; Chan, Wai-Yip
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Illinois Univ., Chicago, IL, USA
Volume :
2
fYear :
1997
fDate :
26-29 Oct 1997
Firstpage :
799
Abstract :
Joint coding of motion information and motion compensation residual information in a rate-distortion optimisation setting can entail a prohibitive computational cost. For a multiresolution motion compensation framework, we propose a fast and efficient scheme that finds the displacement vectors while accounting for the rate-distortion performance of residual coding. The proposed scheme is compared with a popular H.263 test model, with its advanced prediction mode enabled. For various common QCIF video test sequences, the proposed scheme outperforms H.263 by between 0.5-1.1 dB in PSNR, while using roughly a 15% higher complexity than H.263. The similarity of our framework to H.263 renders our results useful for understanding the performance complexity trade-offs of encoders for the current H.263 standard and its descendents
Keywords :
computational complexity; image resolution; image sequences; motion compensation; optimisation; prediction theory; rate distortion theory; transform coding; video coding; H.263 standard; H.263 test model; PSNR; QCIF video test sequences; computational cost; displacement vectors; joint coding; motion compensation residual information; multiresolution motion compensation; multiresolution motion field; performance complexity trade-offs; prediction mode; prediction residual; rate-distortion optimisation; rate-distortion performance; reduced-complexity rate-distortion optimization; residual coding; transform coding; Bit rate; Computational efficiency; Lagrangian functions; Motion compensation; PSNR; Predictive models; Rate-distortion; Redundancy; Signal resolution; Spatial resolution; Testing; Voltage control;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image Processing, 1997. Proceedings., International Conference on
Conference_Location :
Santa Barbara, CA
Print_ISBN :
0-8186-8183-7
Type :
conf
DOI :
10.1109/ICIP.1997.638617
Filename :
638617
Link To Document :
بازگشت