DocumentCode
1717285
Title
MPSoC architecture for H.264/AVC intra prediction chain on SoCLiB platform and FPGA technology
Author
Belhadj, N. ; Turki, M. ; Marrakchi, Z. ; Ben Ayed, M. Ali ; Masmoudi, N. ; Mehrez, H.
Author_Institution
Nat. Eng. Sch. of Sfax, Univ. of Sfax, Sfax, Tunisia
fYear
2013
Firstpage
216
Lastpage
219
Abstract
Multiprocessor System on Chip (MPSoC) technology can present an interesting solution to reduce the computational time of complex applications. Execute the H.264/AVC encoder on MPSoC architecture, is becoming an interesting point of research that can mitigate its algorithmic complexity and to resolve the real time constraints. In this paper, we present an efficient MPSoC architecture for the intra prediction process which is an important module of the H.264/AVC video encoder, using Data Level Parallelism (DLP) partitioning. This architecture is tested on an open platform for MPSoC architectures virtual designing (SoCLiB), and validated on FPGA technology. Experimental results show a gain of 74% in term of encoding speed when using four processors for coding a High Definition Video sequence (HDV) compared to uni-processor architecture.
Keywords
computational complexity; data compression; field programmable gate arrays; image sequences; multiprocessing systems; system-on-chip; video coding; DLP partitioning; FPGA technology; H.264/AVC encoder; H.264/AVC intra prediction chain; HDV; MPSoC architecture; SoCLib platform; algorithmic complexity; computational time reduction; data level parallelism; encoding speed; high definition video sequence; multiprocessor system on chip; real time constraints; uni-processor architecture; video encoder; Computer architecture; Encoding; Field programmable gate arrays; Parallel processing; Partitioning algorithms; Program processors; Video coding; FPGA; H.264/AVC; MPSoC; SoCLiB;
fLanguage
English
Publisher
ieee
Conference_Titel
Sciences and Techniques of Automatic Control and Computer Engineering (STA), 2013 14th International Conference on
Conference_Location
Sousse
Print_ISBN
978-1-4799-2953-5
Type
conf
DOI
10.1109/STA.2013.6783133
Filename
6783133
Link To Document