Chapter 5 - Warping-Based Motion Estimation Techniques

Abstract

The performance of the motion estimation and compensation processes can be improved by using advanced motion estimation techniques such as warping-based motion estimation. In motion estimation and compensation using higher order models, a 2-D mesh is used to divide the current frame into nonoverlapping polygonal patches (or elements). The points shared by the vertices of the patches are referred to as “grid points” or “node points.” In the design of a warping-based technique, the most widely used shapes of patches are triangles or quadrilaterals. The mesh structure can be fixed or adaptive. A fixed mesh is one that is built according to a predetermined pattern, e.g., a regular mesh with square patches. An adaptive mesh, on the other hand, is one that is adaptively built according to frame contents and motion. There are two main methods for estimating the motion of the common vertices between adjacent patches: the continuous method and the discontinuous method. The discontinuous method is more flexible and can compensate for more general complex motion, but it generates more motion overhead (four motion vectors per patch) compared to the continuous method (about one motion vector per patch). The chapter investigates the performance of warping-based methods and compares it to that of block-matching methods. The algorithms implemented in this study were block-matching algorithm (BMA), BMA-HO, and warping-based algorithm.