A fast and adaptive motion tracking scheme
Xuedong Liu, Yihua Tan, Liu Jian · 2008
The unacceptable computational cost of motion tracking (MT) using full search (FS) led to explosive research in MT. A fast and adaptive MT scheme is presented in this paper. The correlation of spatial and temporal neighboring blocks is used firstly to predict an initial motion vector (MV) for the current block. By the region motion intensity, various search patterns are chosen adaptively to refine the MV. To avoid being trapped in unnecessary search, the halfway stop technique is adopted. Combined with the check priority rule, the search is sped up further. Experimental results show the proposed scheme provides the competitive performance with lower computational complexity.