DocumentCode :
2086401
Title :
A fast mode decision algorithm of 4x4 block intra prediction for H.264/AVC
Author :
Cheng, Yun ; Wu, Min
Author_Institution :
Sch. of Inf. Sci. & Eng., Central South Univ., Changsha, China
Volume :
1
fYear :
2008
fDate :
17-19 Nov. 2008
Firstpage :
1231
Lastpage :
1236
Abstract :
It is well known that H.264/AVC has great advantage of coding efficiency compared with the successful prior coding standards and it can save about 50% bit-rate under the same reconstructed picture quality. But the complexity of H.264/AVC encoder is also very high and it limits the application of H.264/AVC in the domain of real time video communication. Based on some observations in the experiments, we propose a novel fast mode decision algorithm of 4×4 block intra prediction for H.264/AVC in this paper. The proposed algorithm utilizes the following facts: (i) the horizontal prediction mode has higher possibility to be the best prediction mode than the other eight modes, (ii) The RD costs between the best prediction mode and its adjacent prediction mode have strong correlation, (iii) some 4×4 blocks¿ coefficients of DCT are all zeroes in some cases. Therefore, it can reduce the computational complexity of the H.264/AVC encoder remarkably. Experimental results show that the reduced time of our proposed algorithm is about 45% less than that of H.264/AVC, at the expense of 2% or so bit-rate increasing.
Keywords :
computational complexity; discrete cosine transforms; video coding; DCT; H.264-AVC encoder; advanced video coding; block intra prediction; computational complexity; fast mode decision algorithm; real time video communication; Automatic voltage control; Computational complexity; Costs; Decoding; Discrete cosine transforms; Image reconstruction; Intelligent systems; Knowledge engineering; MPEG 4 Standard; Video coding;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent System and Knowledge Engineering, 2008. ISKE 2008. 3rd International Conference on
Conference_Location :
Xiamen
Print_ISBN :
978-1-4244-2196-1
Electronic_ISBN :
978-1-4244-2197-8
Type :
conf
DOI :
10.1109/ISKE.2008.4731118
Filename :
4731118
Link To Document :
بازگشت