3D Reconstruction of human bones based on dictionary learning. (November 2017)

• Record Type: Journal Article
• Authors: Zhang, Binkai; Wang, Xiang; Liang, Xiao; Zheng, Jinjin
• Is Part Of: Medical engineering & physics. Volume 49(2017)
• Page Start: 163
• Page End: 170
• Publication Date: 2017-11
• Subjects: 3D reconstruction -- Dictionary learning -- CT image data -- Triangular mesh -- Bio-CAD; Biomedical engineering -- Periodicals; Biomedical Engineering -- Periodicals; Physics -- Periodicals; Génie biomédical -- Périodiques; Biomedical engineering; Electronic journals; Periodicals; 610.28
• DOI: 10.1016/j.medengphy.2017.07.012
• Languages: English
• ISSNs: 1350-4533

Abstract

Highlights: We present an effective 3D reconstruction method of human bones from sequential CT images based on dictionary learning. The objective function containing approximation term and regularization term is proposed to improve accuracy and regularization of the reconstructed mesh. To fit different reconstructing conditions, we use a balance coefficient K and propose a preferential selection method to optimize the approximation term and the regularization term. Dictionary updating and sparse coding are iterated alternatively to complete the optimization within our reconstruction framework. Experimental results show that the proposed method can obtain high-precision and high-quality triangular mesh model. Abstract: An effective method for reconstructing a 3D model of human bones from computed tomography (CT) image data based on dictionary learning is proposed. In this study, the dictionary comprises the vertices of triangular meshes, and the sparse coefficient matrix indicates the connectivity information. For better reconstruction performance, we proposed a balance coefficient between the approximation and regularisation terms and a method for optimisation. Moreover, we applied a local updating strategy and a mesh-optimisation method to update the dictionary and the sparse matrix, respectively. The two updating steps are iterated alternately until the objective function converges. Thus, a reconstructed mesh could be obtained with high accuracy and regularisation. The experimental results show that the proposed method has the potential to obtain high precision and high-quality triangular meshes for rapid prototyping, medical diagnosis, and tissue engineering.