General Object Reconstruction based on Simplex Meshes
Abstract
In this paper, we propose a general tridimensional reconstruction algorithm of range and volumetric images, based on deformable simplex meshes. The algorithm is able to reconstruct surfaces without any restriction on their shape or topology. The different tasks performed during the reconstruction include the segmentation of objects in the scene, the extrapolation of missing data and the control of smoothness, density and geometric quality of the reconstructed model. All surfaces are represented as simplex meshes, that are unstructured meshes whose topology is dual of triangulations. The reconstruction takes place in two stages. First, we initialize the model either manually or using an automatic initialization routine. After the first fit, the topology of the model can be modified by creating holes or increasing its genus. Finally, an iterative adaptation or refinement algorithm decrease the distance of the model from the data while preserving a high geometric and topological quality. We have applied our algorithm to several medical images or range images.