21st International Conference on VLSI Design (VLSI Design 2008)
Low Power Hardware Architecture for VBSME Using Pixel Truncation
Hyderabad, India
January 04-January 08
ISBN: 0-7695-3083-4
This paper presents an efficient architecture to imple- ment low power variable block size motion estima- tion (VBSME) using full search. Power reduction is achieved by performing the search in two steps: low pixel resolution and full pixel resolution. We analysed the computation and memory units needed to support these two search modes. The proposed architecture re- duces the total energy consumption by 50% with 6% additional area compared to the conventional architec- ture.
Citation:
Asral Bahari, Tughrul Arslan, Ahmet T. Erdogan, "Low Power Hardware Architecture for VBSME Using Pixel Truncation," vlsid, pp.389-395, 21st International Conference on VLSI Design (VLSI Design 2008), 2008