Entropic Method for 3D Point Cloud Simplification

Abstract

To represent the surface of complex objects, the samples resulting from their digitization can contain a very large number of points. Simplification techniques analyse the relevance of the data. These simplification techniques provide models with fewer points than the original ones. Whereas reconstruction of a surface, with simplified point cloud, must be close to the original. In this article, we develop a method of simplification based on the concept of entropy, which is a mathematical function that intuitively corresponds to the amount of information this allows considering only relevant points.