Construction method of vertebral body geometric shape statistical model fused curvature features

融合曲率特征的椎体几何形态统计模型构建方法

¹ Hanshan Normal University, Chaozhou 521041, China
² Yuebei People's Hospital, Shaoguan 510525, China
³ Artificial Intelligence and Big Data College, Chongqing College of Electronic Engineering, Chongqing 401331, China
⁴ School of Computer Science, China West Normal University, Nanchong 637002, China
⁵ School of Software, Northwestern Polytechnical University, Xi’an 710072, China

Received: 18 September 2021

Abstract

现有医学样本库缺乏不同年龄段脊椎的几何形态统计模型, 因此建立一个精度较高、定位准确的椎体几何形态统计模型库就显得迫在眉睫。提出一种基于融合曲率特征的椎体三维几何统计模型构建方法, 根据曲率特征对人体腰椎三维重建模型进行识别和定位, 输出各模型的样本矩阵; 用优化的ICP算法对样本矩阵进行对齐与配准; 用PCA对样本模型进行配准训练, 组成统计模型样本库。

摘要

At present, the existing medical sample database lacks geometric statistical models of vertebrae of different ages. Therefore, it is essential to establish a geometric statistical model database of vertebrae with high precision and accurate positioning. This paper proposes a construction method of three-dimensional geometric statistical model of vertebrae based on fused curvature features. First, we identify and locate the three-dimensional reconstruction model of the human lumbar spine according to the curvature characteristics, and output the sample matrix of each model; then, we use the optimized ICP algorithm to align and register the sample matrix; finally, we apply PCA to perform registration training on the sample model, and form the sample database of statistical model in the end.