In this paper, we propose a graphics processing unit (GPU)-based motion estimation for H.264/AVC by rearranging the encoding order of 4times4 blocks. Previous research by Kelly and Kokaram (2003) and Lin et al. (2006) assume no dependency among adjacent blocks, which is not true for H.264/AVC. The proposed method not only overcomes the dependency problem, but also can be used for both full search and fast search. Experimental results show that our implementation is forty five faster than our SDVTD optimized CPU implementation.
Published in:
Audio, Language and Image Processing, 2008. ICALIP 2008. International Conference on
Date of Conference: 7-9 July 2008