| 摘要 |
A method and system is presented for generating a full motion field of an object, when performing non-rigid image registration between a first and a second image of the object. A mesh grid generator generates in the first image a mesh grid having a plurality of mesh nodes, for each of a hierarchy of mesh resolution levels. A nodal motion estimator estimates, for each mesh node, at least one nodal motion vector that describes a matching of the mesh node with a corresponding mesh node in the second image. At each mesh node, multi-level block matching is performed, using a similarity measure based on pattern intensity. A smoothness constraint is imposed, in order to reconstruct the nodal motion vector for those mesh nodes at which mismatching occurs. A motion field interpolator determines a local motion vector for any desired point of interest within the first image, by interpolating from the nodal motion vectors of the surrounding mesh nodes. |
