Abstract:
This paper details the development of finite element models (FEMs) for analyzing the biomechanics of human joints, highlighting a semi-automated method to address the challenges of using high-resolution medical data. The authors, P-L. Bossart and K. Hollerbach, explain the sequential process for creating these models, beginning with the acquisition of high-resolution data from an industrial X-ray CT scanner, which provides the precise geometric detail necessary for defining articular surfaces. The data then undergoes segmentation, a process made difficult by the inhomogeneous nature of trabecular bone. To overcome this, the researchers use a simple bone attenuation model, morphological reconstruction, and watershed lines to automate coarse segmentation, followed by human interaction to refine the results. Next, polygonal surfaces are extracted from the segmented data, and a decimation algorithm is applied to reduce the number of vertices, making the data more manageable. A key innovation is the use of a template-based approach to generate hexahedral volumetric meshes, which are preferred for dynamic simulations due to their faster convergence compared to tetrahedral meshes. This method deforms a pre-defined template to fit the specific geometry of a bone, which helps to minimize development time for patient-specific models. Finally, these models, which include bone and soft tissue structures like ligaments and cartilage, are used with the NIKE3D finite element code to simulate joint motion and calculate soft tissue stresses, providing tools for biomechanics researchers and clinicians.
