Modeling and characterizing collagen fiber bundles

Collagen fibers are prominent constituents of many soft tissues. The fibers are able to carry load along their fiber axis, and, therefore, their structural arrangement renders the mechanical behavior of biological tissue anisotropic. In some tissues, e.g., blood vessels, tendons and ligaments, the fibers form fiber bundles. Structure-based biomechanical models of soft tissues strongly rely on a clear understanding of the underlying collagen fiber bundle structure and require precise quantitative structural information. The paper presents a novel fiber bundle model that satisfactorily explains the observed bundle structure and reveals meaningful parameters for bundle characterization. The model is capable of describing a wide variety of natural bundle structures. It is shown that the identified bundle parameters can be extracted from real histological images using computer vision algorithms. The generative nature of the model allows the creation of artificial fiber bundle images with characteristics very similar to real images.