Shape modeling with spline partitions

Shape modelling (with methods that output shapes) is a new and important task in Bayesian nonparametrics bioinformatics. In this work, we focus on nonparametric for capturing shapes by partitioning space using curves. related the classical Mondrian process used to partition spaces recursively with axis-aligned cuts, widely lied multi-dimensional relational data. The outputs hyper-rectangles. Recently, random tessellation was introduced as generalization of process, domain non-axis aligned cuts an arbitrary dimensional space, outputting polytopes. Motivated these processes, propose novel parallelized approach curves, enabling complex data-shapes be acquired. We apply our method HIV-1-infected human macrophage image dataset, also simulated datasets sets illustrate approach. compare support vector machines, forests state-of-the-art computer vision such simple linear iterative clustering super pixel segmentation. develop R package available at https://github.com/ShufeiGe/Shape-Modeling-with-Spline-Partitions .