A local geometric preserving approach for interior deformation fields measurement from MR volumetric images of human tissues

摘要

While a variety of different features matching algorithms have been reported in rigid areas, few features matching algorithm used in non-rigid area have been reported. This work is concerned about interior deformation fields measurement of non-rigid, non-uniform human tissue or organs from 3D magnetic resonance volumetric images. In this paper, a local geometric preserving approach was proposed to find homologous features from a given features in MR volume obtained at the initial state. Three dimension invariant moment and geometric preserving property of a local 3D region have been used to design the function which was used to measure the strength of match of the candidate feature pair. Interior density deformation fields is then inferred to use a linear approximate approach in an irregular tetrahedra finite element model. To test the validity of the proposed approach, it is applied to actual MR volumetric images obtained from a volunteer's finger. The primary result is consistent with the fact.