Sommer H J, Miller N R, Pijanowski G J
J Biomech. 1982;15(3):171-80. doi: 10.1016/0021-9290(82)90249-4.
Many analytical biomechanics methods require extensive three-dimensional descriptions of anatomical geometry. In particular, researchers requiring the three-dimensional coordinates of specific boney landmarks (e.g. tendon and ligament attachments) are often forced to extrapolate such measurements from an experimental specimen set to their subject geometry. This work offers an approach to two problems inherent above; accurate extrapolation of specimen landmark locations to subject homologues and statistical accumulation of normative three-dimensional anatomical landmark data bases. A least squares solution for an affine scaling transformation from specimen to subject is used which incorporates both right-left and same hand comparisons. A two stage technique is formulated to consecutively remove landmark location variation and to size a normative specimen from a set of similar specimens. This ability to statistically represent a specimen set will provide better geometric models for other analytical studies and prosthetic design and evaluation.
许多分析生物力学方法需要对解剖几何结构进行广泛的三维描述。特别是，需要特定骨标志（如肌腱和韧带附着点）三维坐标的研究人员，常常被迫将此类测量值从实验标本集外推到其研究对象的几何结构上。这项工作为上述两个固有问题提供了一种方法；将标本标志位置准确外推到研究对象的同源物上，以及规范三维解剖标志数据库的统计积累。使用了一种从标本到研究对象的仿射缩放变换的最小二乘解，该解结合了左右和同手比较。制定了一种两阶段技术，以连续消除标志位置变化，并从一组相似标本中确定一个规范标本的大小。这种对标本集进行统计表示的能力将为其他分析研究以及假体设计和评估提供更好的几何模型。
