Multi-resolution surface modeling from multiple range views

Abstract

This paper presents a multi-resolution surface modeling technique. Several registered range views obtainedfrom different viewpoints are first integrated into a non-redundant surface triangulation. The integration technique is based on the reparametrization of the canonic subsets of the Venn diagram of the set of views. The resulting triangulation is then input to a sequential optimization process that computes different levels of resolution of the surfaces of interest.

Description of Complex Objects from Multiple Range Images Using an Inflating Balloon Model

“…As stated earlier, our goal is to achieve a correct mapping of a triangulated mesh to the surface of an object, hence there are still many ways to improve the resulting model we have achieved, including using the algorithms presented in [2] to improve triangle fitting errors, or the method in [10] to merge small triangles into larger ones without affecting the fitting error for constructing a hierarchical representation. Local smooth patches can also be constructed for high level surface property analysis.…”

“…Very few researchers (e.g. [10]) try to use multiple densely sampled range images. There are two difficulties in using these range images.…”

Principal component analysis with missing data and its application to object modeling

“…Ahuja and Veenstra [ I ] constructed an octree object model from orthogonal views. Soucy and Laurendeau [ 16] proposed to triangulate each view and merge multiple views via a Venn diagram when the transformation is known. Because building a Venn diagram is combinatorial, only four-view merging is pre-2 sented in their work.…”

Real-time 3D model acquisition

“…Previous implementations of this pipeline have been proposed, with a variety of technologies at each of the stages [Soucy and Laurendeau 1992;Turk and Levoy 1994;Curless and Levoy 1996]. Although such systems have been used to create complex, high-resolution models, as in several recent sculpture digitization projects [Rushmeier et al 1998;Levoy et al 2000;Miyazaki et al 2000], doing so is neither fast nor easy.…”