Abstract:
This paper presents a novel computational framework for generating realistic muscle fibre architectures within arbitrary muscle volumes, facilitating biomechanical analysis and interactive visualization. Starting from a triangular surface mesh of a muscle with expert-defined origin and insertion landmarks, the method constructs a unit fibre template based on Bézier curves and Sobol sampling, which is then affinely fitted and morphologically mapped into the muscle volume using slice-by-slice barycentric-coordinate interpolation. To ensure anatomical fidelity, fibres near extensive attachment areas are trimmed and reconstructed, while iterative smoothing enhances path realism. Implemented in C++ within the Multimod Application Framework and leveraging VTK, the algorithm decomposes muscle volumes into hundreds of fibre polylines in under 1000 ms, enabling real-time interaction and rapid updates during simulated motions. Qualitative comparisons with established anatomical atlases and the commercial Truegrid demonstrate close visual agreement, suggesting that the generated fibre paths can underpin biomechanically relevant metrics such as lever arm estimation and force–length–velocity relationships. This high-performance muscle modelling technique holds promise for educational platforms and future clinical biomechanics applications, offering a balance between computational speed and anatomical accuracy.
