Statistic Model of the Spine in Three-Dimension Geometry

Abstract

The study of the statistic model of the spine in three-dimension (3-D) geometry aims to provide a scientific basis for the spine and vertebra related medical surgery. In this paper, we adopt the Active Sharpe Model (ASM) to build a spinal statistical model. That is, we first locate and mark the feature points of the three-dimensional reconstructed medical Computed Tomography images, so as to obtain the shape matrix of each spine sample. Second, we align and register the shape matrix in the sample set with Iterative Closest Point (ICP). Third, we train the samples with Principal Component Analysis (PCA) and build the spinal statistical model in 3-D geometry. Finally, we evaluate the proposed model.