Title:ToSkA: Topological Skeleton Analysis for Network-Based Shape Representation and Evaluation of Objects from Cells to Death Stars

Abstract:Shape analysis and classification are popular methods for biologists, biophysicists and mathematicians investigating relationships between object function and form. Classic shape descriptors, such as sphericity, can be powerful but may be insufficient for more complex shapes. Here, we present 'napari-toska' a topological skeleton based method to analyze complex objects by representing their shape asymmetries as networks. Using global neighborhood principles, classic network science metrics and spatial feature embedding we create instance segmentation object profiles to be used for immediate or downstream classification. napari-toska can also follow temporal dynamics and identify network features capable of differentiating between experimental phenotypes. We incorporated the capacity to measure absolute spatial features of objects to bring in aspects of scale. Furthermore, napari-toska identifies certain segmentation errors through the emergence or loss of network cycles. Combined, napari-toska functions allow for flexible and in-depth shape profiling of intricate shapes often observed in biological and physical settings where robust, yet precise, system configuration is essential to downstream processes.