Fast sub-pixel motion estimation having lower complexity

This paper presents fast sub-pixel motion estimation having less computational complexity. The proposed methods are based on mathematical models of the motion compensated errors for moving picture compression. The methods need no interpolation from the integer-pixel search results. In order to decide the coefficients of the models, the motion compensated errors of the neighboring pixels around motion vectors are used. The errors were already obtained during the pixel accuracy full search. The models estimate the motion compensated errors at sub-pixel accuracy level. The proposed mathematical models are referred to as Model 1, Model 2, Model 3, modified Model 2, and modified Model 3, respectively. The simulation results show that the performance depends on the accuracy of the models. The performance of the proposed methods, having less computational complexity, is close to that of the full search method.