一种高度自动化的三维CT数据腰椎分割方法

A Highly Automatic Lumbar Vertebrae Segmentation Method Using 3D CT Data

针对现有分割方法,分别存在无法完全分割、不能适应所有数据或过多的手动干预等局限性问题,提出了一种用于CT图像单节腰椎分割的新方法.首先从CT图像中分割整条脊椎,然后再分别断开单个椎体和椎小关节,从而使待分割的一节腰椎与其他部位分离.利用人造数据和真实数据对本方法进行了评估,结果显示本法可以高度自动化地分割任意单节腰椎.本方法是医生进行辅助诊断的重要方法,在相关领域的研究中具有广泛意义.

Computer image processing technology is commonly used in clinical diagnoses of lumbar vertebrae problems and in medical research to segment single lumbar vertebrae with 3D CT images.Problems with existing segmentation methods include the inability to completely segment,failure to process all the data,or an excessive need for hand labor.A new method was proposed to segment individual lumber vertebrae using CT images.The entire backbone is first segmented from CT images,and then each single centrum and intervertebral facet joints are disconnected,thereby segmenting out the lumbar vertebra of interest.Evaluation using real data and synthetic data indicate a successful highly automatic segmentation of single lumbar vertebrae.This method provides doctors with an important means of computer-aided diagnosis,as well as being useful in biomechanics and other fields of research.