Title :
A fast motion estimation algorithm equivalent to exhaustive search
Author :
Gharavi-Alkhansari, Mohammad
Author_Institution :
Tarbiat Modares Univ., Tehran, Iran
Abstract :
A fast algorithm is proposed for block motion estimation for video sequences. The proposed algorithm is proven to be equivalent to exhaustive search. In a multiresolution approach, it uses a mathematically derived threshold to prune search candidates whose low-resolution versions are too far from the low resolution version of the block for which a best match is sought. Experimental results show that speed ups of around 36, compared to full search, may be achieved, for some typical test video sequences. This is the fastest full-search-equivalent motion estimation reported in the literature to date, and has speed ups comparable to inexact fast motion estimation methods
Keywords :
image resolution; image sequences; mean square error methods; motion estimation; search problems; video signal processing; MSE; exhaustive search; fast motion estimation algorithm; mean absolute difference; mean squared error; multiresolution approach; test video sequences; video sequences; Computational efficiency; Decoding; Motion estimation; Road transportation; Search methods; Testing; Uninterruptible power systems; Video coding; Video compression; Video sequences;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2001. Proceedings. (ICASSP '01). 2001 IEEE International Conference on
Conference_Location :
Salt Lake City, UT
Print_ISBN :
0-7803-7041-4
DOI :
10.1109/ICASSP.2001.941139