Fibre topography of the extracellular matrix governs local mechanical properties and cellular behaviour including migration and gene expression. While quantifying properties of the fibrous network provides valuable data that could be used across a breadth of biomedical disciplines, most available techniques are limited to two dimensions and, therefore, do not fully capture the architecture of three-dimensional (3D) tissue. The currently available 3D techniques have limited accuracy and applicability and many are restricted to a specific imaging modality. To address this need, we developed a novel fibre analysis algorithm capable of determining fibre orientation, fibre diameter and fibre branching on a voxel-wise basis in image stacks with distinct fibre populations. The accuracy of the technique is demonstrated on computer-generated phantom image stacks spanning a range of features and complexities, as well as on two-photon microscopy image stacks of elastic fibres in bovine tendon and dermis. Additionally, we propose a measure of axial spherical variance which can be used to define the degree of fibre alignment in a distribution of 3D orientations. This method provides a useful tool to quantify orientation distributions and variance on image stacks with distinguishable fibres or fibre-like structures.

中文翻译：

---

纤维网络分段图像堆栈中纤维取向、直径和分支的三维计算

细胞外基质的纤维形貌控制着局部机械特性和细胞行为，包括迁移和基因表达。虽然量化纤维网络的特性提供了可在广泛的生物医学学科中使用的有价值的数据，但大多数可用的技术仅限于二维，因此不能完全捕获三维 (3D) 组织的结构。目前可用的 3D 技术的准确性和适用性有限，并且许多技术仅限于特定的成像模式。为了满足这一需求，我们开发了一种新颖的纤维分析算法，能够在具有不同纤维群的图像堆栈中以体素为基础确定纤维方向、纤维直径和纤维分支。该技术的准确性在计算机生成的涵盖一系列特征和复杂性的幻影图像堆栈以及牛腱和真皮中弹性纤维的双光子显微镜图像堆栈上得到了证明。此外，我们提出了一种轴向球面方差的测量方法，可用于定义 3D 方向分布中的纤维排列程度。该方法提供了一种有用的工具来量化具有可区分的纤维或纤维状结构的图像堆栈上的方向分布和方差。