DocumentCode :
2472011
Title :
A fast intra prediction mode decision method based on high correlation of video sequences
Author :
Park, Jihan ; Jeon, Yongwook ; Har, Dongsoo
Author_Institution :
Dept. of Inf. & Commun., Gwangju Inst. of Sci. & Technol., Gwangju, South Korea
fYear :
2010
fDate :
14-17 March 2010
Firstpage :
157
Lastpage :
161
Abstract :
H.264/AVC achieved better coding efficiency and visual quality than the previous video coding standards such as H.262, MPEG-1, MPEG-2, H.263 and MPEG-4 part2 at the same bit rate. To achieve the better coding efficiency, H.264/AVC uses the rate distortion optimization technique (RDO). The computational complexity is increased by using RDO calculation in all possible modes of intra prediction. To reduce this complexity, a number of fast algorithms for inter/intra-predications are proposed. One of approaches to reducing complexity is based on the minimization of the candidate modes by using preprocessing. In this paper, we propose a fast intra prediction mode decision method using high correlation in video sequences. The proposed method decides intra prediction mode using high correlation of the video sequence. The experimental results show that the encoding time saving can be achieved up to maximally 59% with negligible PSNR (Peak Signal-to-Noise Ratio) drop and a slight increment in bit rate compared to the results of the joint model (JM) reference software of H.264/AVC proceedings.
Keywords :
computational complexity; decision theory; image sequences; prediction theory; rate distortion theory; video coding; H.264/AVC; candidate mode minimization; coding efficiency; computational complexity; fast intra prediction mode decision method; peak signal-to-noise ratio drop; rate distortion optimization technique; video coding standards; video sequences; visual quality; Automatic voltage control; Bit rate; Computational complexity; MPEG 4 Standard; Rate-distortion; Redundancy; Transform coding; Video coding; Video compression; Video sequences;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Industrial Technology (ICIT), 2010 IEEE International Conference on
Conference_Location :
Vi a del Mar
Print_ISBN :
978-1-4244-5695-6
Electronic_ISBN :
978-1-4244-5696-3
Type :
conf
DOI :
10.1109/ICIT.2010.5472671
Filename :
5472671
Link To Document :
بازگشت