Open AccessWatertight 2-manifold 3D bone surface model reconstruction from CT images based on visual hyper-spherical mapping
Author(s) -
Tianran Yuan,
Hongsheng Zhang,
Hao Líu,
Juan Du,
Huan Yu,
Yimin Wang,
Yabin Xu
Publication year - 2021
Publication title -
mathematical biosciences and engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.451
H-Index - 45
eISSN - 1551-0018
pISSN - 1547-1063
DOI - 10.3934/mbe.2021068
Subject(s) - thresholding , surface reconstruction , surface (topology) , computer vision , segmentation , artificial intelligence , projection (relational algebra) , computer science , sampling (signal processing) , 3d reconstruction , volume (thermodynamics) , mathematics , algorithm , image (mathematics) , geometry , physics , filter (signal processing) , quantum mechanics
This paper proposes a general algorithm to reconstruct watertight 2-manifold 3D bone surface model from CT images based on visual hyper-spherical mapping. The reconstruction algorithm includes three main steps: two-step thresholding, initial watertight surface reconstruction and shape optimization. Firstly, volume sampling points of the target bone with given narrower threshold range are extracted by thresholding with combination of 3D morphology operation. Secondly, visible points near the bone's outer surface are extracted from its corresponding volume sampling points by hyper-spherical projection mapping method. Thirdly, implicit surface reconstruction algorithm is employed on the extracted visible surface points to obtain an initial watertight 3D bone surface model which is used as the deformation model in the following accurate bone surface model generation stage. Finally, the initial surface model is deformed according to the segmentation data with wider threshold range under given constraints in order to achieve an accurate watertight 3D bone surface model. Experiment and comparison results show that the proposed algorithm can reconstruct watertight 3D bone surface model from CT images, and local details of the bone surface can be restored accurately for the cases used in this paper.