An Implementation of Data Reusable MPEG Video Coding Scheme

Authors: Vasily G. Moshnyaga

Abstract:

This paper presents an optimized MPEG2 video codec implementation, which drastically reduces the number of computations and memory accesses required for video compression. Unlike traditional scheme, we reuse data stored in frame memory to omit unnecessary coding operations and memory read/writes for unchanged macroblocks. Due to dynamic memory sharing among reference frames, data-driven macroblock characterization and selective macroblock processing, we perform less than 15% of the total operations required by a conventional coder while maintaining high picture quality.

Keywords: Data reuse, adaptive processing, video coding, MPEG

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1079264

References:

[1] M. Takahashi, et al, A 60-MHz240mW MPEG-4 Videophone LSI with 16Mb Embedded DRAM, IEEE JSSC, 35 (11), pp.1713-1721, Nov.2000.
[2] T. Hashimoto, et al, A 90mW MPEG4 Video Codec LSI with the Capability for Core Profile, IEEE ISSCC 2001, pp.140-141.
[3] J. Tuan, et al, On the data reuse and memory bandwidth analysis for full-search block-matching VLSI architecture, IEEE Trans. CASVT, 12 (1) pp.61-72, 2002
[4] J. Davidson, et al, Memory access coalescing: A technique for eliminating redundant memory accesses, PLDI, 29 (6) 1994.
[5] D.Kolson, et al, Integrating program transformations in the memory based synthesis of image and video applications, IEEE ICCAD'94, 27-30.
[6] S. Wuytack, et al, ''Formalized methodology for data reuse exploration for low-power hierarchical memory mappings'', IEEE TVLSI Systems, 6 (4) pp.529-537, Dec. 1998..
[7] T. Onoye, et al., A VLSI architecture for MPEG2 real time motion estimator, IEEE ISCAS 1996
[8] V. G. Moshnyaga, Reduction of memory accesses in motion estimation by block-data data reuse, Proc. IEEE ICASSP'02
[9] V.G. Moshnyaga, A new computationally adaptive formulation of block-matching motion estimation, IEEE Trans. CASVT, 11(1), Jan. 2001.