optimization - Registration of 3D surface mesh to 3D image volume -
i have accurate mesh surface model of implant i'd optimally rigidly align computed tomography scan (scalar volume) contains exact same object. i've tried detecting edges in image volume canny filter , doing iterative closest point alignment between edges , vertices of mesh, it's not working. tried voxelizing mesh, , using image volume alignment methods (mattes mutual) yields inconsistent results.
any other suggestions?
thank you.
generally, mesh , volume 2 different data structures. have either convert mesh volume or convert volume mesh.
i recommend doing segmentation of volume data first, segment out issues want register. canny filter might not enough segment border clearly. recommend level-set method , active contour model. these 2 used in medical image processing. these 2 topics, recommend professor chunming li's work.
and after segmentation of volume data, might able reconstruct mesh model of volume marching cubes. vertexes of 2 mesh registered through simple icp algorithm.
however, workaround instead of real registration, takes time segmentation.
Comments
Post a Comment