摘要
该文以人体膝关节个性化三维建模为目的,并论证该模型应用于在体运动测量的可行性。基于可变形统计模型理论,寻找模型表面的特征配对点,建立膝关节统计模型训练集;通过主分量分析的方法确定特征点在三维空间的分布规律;在重建新模型时,通过统计模型变形,以X线图像中膝关节的二维透视信息为形变基准,从而实现膝关节个性化三维建模。研究结果表明:相比MRI(magnetic reso-nance imaging)模型,本文重建的完整膝关节三维骨骼模型精度为0.5mm,该模型用于运动测量的精度为0.59mm和1.25°。可变形统计模型技术可以用于建立受试者膝关节个性化三维模型,并测量膝关节的在体三维运动。
The objective of this study was to reconstruct patient-specific 3D knee joint models and measure the in-vivo knee joint kinematics. The training set for the statistical deformation model was constructed from characteristic points of a group of knee models. Principal component analysis was used to find the spatial variations of the points with a patient-specific 3D knee model developed based on 2D radiographic images and the trained models. The 3D bone model was within 0.5 mm compared to a MRI （magnetic resonance imaging） model. The kinematic measurements were accurate within 0. 59 mm and 1.25°. The statistical deformation model was validated as a useful method for reconstructing patient-specific knee models and measuring knee kinematics in three dimensions.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第1期139-144,共6页
Journal of Tsinghua University(Science and Technology)
基金
国家"九七三"重点基础研究项目(2011CB707701)
高等学校博士学科点专项科研基金资助项目(20090002110023
20110002110024)
关键词
生物力学
可变形统计模型
膝关节
biomechanics
statistical deformation model
knee joint
