Department of Oncology and Metabolism, University of Sheffield, UK; Insigneo Institute for in Silico Medicine, University of Sheffield, UK; Department of Industrial Engineering, School of Engineering and Architecture, Alma Mater Studiorum - Università di Bologna, Bologna, Italy.
Department of Industrial Engineering, School of Engineering and Architecture, Alma Mater Studiorum - Università di Bologna, Bologna, Italy.
J Biomech. 2019 Mar 27;86:232-237. doi: 10.1016/j.jbiomech.2019.01.041. Epub 2019 Jan 30.
Digital Volume Correlation (DVC) is used to measure internal displacements and strains in bone. Recent studies have shown that Synchrotron radiation micro-computed tomography (SR-microCT) can improve the accuracy and precision of DVC. However, only zero-strain or virtually-moved test have been used to quantify the DVC uncertainties, leading to potential underestimation of the measurement errors. In this study, for the first time, the uncertainties of a global DVC approach have been evaluated on repeated SR-microCT scans of bovine cortical bone (voxel size: 1.6 μm), which were virtually deformed for different magnitudes and along different directions. The results showed that systematic and random errors of the normal strain components along the deformation direction were higher than the errors along unstrained directions. The systematic percentage errors were smaller for larger virtual deformations. The random percentage error was in the order of 10% of the virtual deformation. However, higher errors were localized at the boundary of the volumes of interest, perpendicular to the deformation direction. When only the central region of the samples was considered (100 µm layers removed from the borders where the deformation was applied), the errors in the direction of virtual deformation were comparable to the errors in the unstrained directions. In conclusion, the method presented to estimate the uncertainties of DVC is suitable for testing anisotropic specimens as cortical bone. The good agreement between the uncertainties in measurements of strain components obtained with this approach and with the simpler zero-strain-test suggests that the latter is adequate in the tested deformation scenarios.
数字体视学相关技术（DVC）用于测量骨骼内部的位移和应变。最近的研究表明，同步辐射微计算机断层扫描（SR-microCT）可以提高 DVC 的准确性和精度。然而，仅使用零应变或虚拟移动测试来量化 DVC 的不确定性，可能会低估测量误差。在这项研究中，首次在对牛皮质骨（体素大小：1.6μm）的重复 SR-microCT 扫描中评估了全局 DVC 方法的不确定性，这些皮质骨被虚拟变形为不同的大小和不同的方向。结果表明，沿变形方向的法向应变分量的系统和随机误差高于未受应变方向的误差。对于较大的虚拟变形，系统百分比误差较小。随机百分比误差为虚拟变形的 10%左右。然而，更高的误差集中在感兴趣区域的边界处，垂直于变形方向。当仅考虑样本的中心区域（从施加变形的边界去除 100μm 层）时，虚拟变形方向的误差与未受应变方向的误差相当。总之，用于估计 DVC 不确定性的方法适用于测试皮质骨等各向异性样本。该方法测量应变分量的不确定性与更简单的零应变测试之间的良好一致性表明，在后一种情况下，在测试的变形场景中是足够的。
