Simple next search position selection for motion estimation iterative search

摘要

A next position selector for motion estimation iterative search improves the efficiency and quality of the integer-pel motion estimation iterative search by easily and quickly determining the center point of the next search area. By taking a position of the SAD value of the present search area and mapping that position to a center point of the next search area, the process of determining the next search area is very efficient. A local minimum avoidance scheme improves avoids being trapped in a local minimum. By remembering the previous position/direction, the search is able to continue in the correct direction to ensure that the search continues without a back and forth search that is trapped in a local minimum. Applications of this improved motion estimation search include stabilizing an image as well as many other applications where motion vectors are used.