DocumentCode :
435680
Title :
Complexity analysis of intra prediction in H.264/AVC
Author :
Zouch, W. ; Samet, Ahmed ; Ben Ayed, M.A. ; Kossentini, F. ; Masmoudi, N.
Author_Institution :
Lab. of Electron. & Inf. Technol., Nat. Eng. Sch. of Sfax, Tunisia
fYear :
2004
fDate :
6-8 Dec. 2004
Firstpage :
713
Lastpage :
717
Abstract :
In this paper, we study the computational complexity of the intra prediction module for video coding software according to H.264/AVC standard. We analyze separately the complexity of each intra mode for the implementation on TMS320C64 platform. The intra prediction unit consists of predicting the block´s pixels from their neighbors according to different directional modes (horizontal, vertical, ...) and others non-directional (DC, planar). For each mode, we compute the number of the executional sub-units (add, shift, load, store...). Together with the number of necessary cycles to execute every instruction and the functional units that can perform these instructions, we can deduce the complexity of every intra mode. To estimate the total complexity of the intra prediction process, we assume that the encoder tests all intra modes. The total cycle´s account is estimated by multiplying the number of cycles of each mode by the number of blocks in the picture. Our results indicate that our methodology for the complexity analyses of the intra prediction unit of the H.264/AVC provides a good approximation with respect to the experimental results.
Keywords :
computational complexity; prediction theory; video coding; H.264-AVC standard; TMS320C64 platform; computational complexity analysis; directional modes; intra prediction module; video coding software; Algorithm design and analysis; Automatic voltage control; Decoding; High performance computing; ISO standards; Information technology; Satellite broadcasting; Streaming media; Strontium; Video coding;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Microelectronics, 2004. ICM 2004 Proceedings. The 16th International Conference on
Print_ISBN :
0-7803-8656-6
Type :
conf
DOI :
10.1109/ICM.2004.1434766
Filename :
1434766
Link To Document :
بازگشت