Research On 3D Thinning Algorithm Based On Rotation-invariant

3D thinning means that erodes an 3D object from outside to inside and keeps the original topology until only one skeleton is left.The skeleton extracted by 3D thinning algorithm is an important foundation for subsequent image analysis and feature extraction,it is widely used in the field of visual measurement,image analysis,3D reconstruction and medical image processing.3D thinning algorithm should keep the topology and avoid to create crevice,it also need to ensure it's effectiveness that only one pixel thickness can be keep.Considered that in practical,the angle of object is unknown,3D thinning algorithm should have rotation invariance so that can keep the topology do not change in the same object but different angle.The existing thinning algorithms which are based on simple point failed to preserve the connectivity of extracted skeleton.we analysis the reason why others fail to protect the structure and propose a new re-checking procedure to fix that problem,By detecting weather the connectivity of its 26-neighborhood changed after deleting points or not,and restore points to insure we could finally got a complete skeleton.This method can suit on most of the thinning algorithms which are based on simple point and fix the cavities to preserve topology structure.The isotropic templates are proposed in this novel,it can accurate extract skeleton,keep the connectivity and topology.Experimental results show that the algorithm proposed in this novel is able to preserve connectivity and topology structure,and it's result is better than most of the other algorithm which are based on templates.