Logo image
Enhanced downhill simplex search for fast video motion estimation
Conference paper   Peer reviewed

Enhanced downhill simplex search for fast video motion estimation

Hwai-Chung Fei, Chun-Jen Chen and Shang-Hong Lai
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol.3767 LNCS, pp.512-523
2005

Abstract

Block-based motion estimation can be regarded as a function minimization problem in a finite two-dimensional space. Therefore, fast block-based motion estimation can be achieved by using an efficient function minimization algorithm instead of using a predefined search pattern, such as the diamond search. The downhill simplex search algorithm is an efficient derivative-free function minimization algorithm. In this paper, we propose several enhanced schemes to improve the efficiency of applying the downhill simplex search algorithm to motion estimation. The proposed enhanced schemes include a new initialization process, a special rounding method, and an early-stop error function evaluation procedure. Experimental results on several benchmarking videos show superior performance of the proposed algorithm over some existing fast block matching methods. © Springer-Verlag Berlin Heidelberg 2005.

Metrics

1 Record Views

Details

Logo image