Increased geometric data complexity, operations on n-dimensional manifold and non-manifold geometries are the modern challenges of many field like CAD, Biomedical Imaging and Physical Simulations. Therefore there is the need of a representation scheme capable to handle these challenges with ease. For this task we propose LAR (Linear Algebraic Representation) which offers many interesting features such as efficient support for topological queries and construction, data representation via matrices and the opportunity for easy parallel processing. LARLIB.jl is the official Open Source implementation of LAR which enables the user to use the features of LAR with operations like: incidences and adjacencies between topological entities, boundary and co-boundaries of Cellular Complexes, Boolean Operations such as union of several complexes. This was possible thanks to the Julia ecosystem that provides and easy environment for scientific computation such as parallel capabilities, definition and operations on CSC Sparse Matrices builtin types, and third-party scientific libraries (IntervalTrees.jl and NearestNeighbors.jl). We will demonstrate the library efficiency through the union of big random cellular complexes and a real Bioinformatics use case with the creation of a compact model from MRI images.
I have a B.Sc. in Computer Engineering at the University of Roma Tre which is the host of the Computational Visual Design Lab (CVDLAB). I worked there as a research fellow for more than a year. During this period I developed LARLIB.jl with Francesco Furiani and Prof. Alberto Paoluzzi.
