Abstract:
Rapid generation of lower limb bone models is essential for clinically applicable patient-specific gait modeling. Simulations not only need bone geometry and pose, but also their muscle attachment sites. Accurate bone geometry and attachment sites are usually labouriously segmented from medical images, which are expensive and inconvenient to obtain. Motion-capture is a routine part of gait assessment but contains relatively sparse geometry information. We present a workflow, implemented in an easy-to-use software application, which uses a statistical model to accurately estimate lower limb bone geometry, pose, and muscle attachment sites from seven commonly used motion-capture landmarks. Our method can significantly reduce modelling time and increase the feasibility of clinical gait modelling.