DocumentCode
2933729
Title
More scanning patterns for entropy coding for H.264
Author
Tao, Yang ; Peng, Yuxing ; Liu, Zhiming
Author_Institution
Nat. Univ. of Defense Technol., Changsha
fYear
2007
fDate
Nov. 28 2007-Dec. 1 2007
Firstpage
490
Lastpage
493
Abstract
In order to reduce the bit rate and to improve the compression performance in video coding in H.264, this paper presents more scanning patterns for entropy coding. Firstly, we investigate the two designs of entropy coding, termed CA VLC and CABAC. And then according to the DCT transform, we deduce that it is possible to adopt more scanning patterns in entropy coding in H.264 to reduce the coding gain at the cost of bit rate. So this paper deduces some different features of an original or residual block, and proposes eight scanning patterns for entropy coding and the corresponding pattern decision method. In the pattern decision method, we make a full search according to the bit rate cost to select the best pattern out of all scanning patterns. Experimental results based on the H.264 JIT reference software of JM 10.1 show that employing the proposed scanning patterns bit rate reduction by up to 4~7% will be obtained with neglectable computation complexity increased and a better performance is also obtained.
Keywords
adaptive codes; block codes; data compression; discrete cosine transforms; entropy codes; variable length codes; video coding; DCT; H.264; adaptive code; block code; data compression; entropy coding; pattern decision method; scanning pattern; variable-length code; video coding; Arithmetic; Automatic voltage control; Bit rate; Computer science; Costs; Discrete cosine transforms; Entropy coding; MPEG 4 Standard; Signal processing algorithms; Video coding;
fLanguage
English
Publisher
ieee
Conference_Titel
Intelligent Signal Processing and Communication Systems, 2007. ISPACS 2007. International Symposium on
Conference_Location
Xiamen
Print_ISBN
978-1-4244-1447-5
Electronic_ISBN
978-1-4244-1447-5
Type
conf
DOI
10.1109/ISPACS.2007.4445931
Filename
4445931
Link To Document