Sylvain Petitjean

Least squares conformal maps for automatic texture atlas generation

A Texture Atlas is an efficient color representation for 3D Paint Systems. The model to be textured is decomposed into charts homeomorphic to discs, each chart is parameterized, and the unfolded charts are packed in texture space. Existing texture atlas methods for triangulated surfaces suffer from several limitations, requiring them to generate a large number of small charts with simple borders. The discontinuities between the charts cause artifacts, and make it difficult to paint large areas with regular patterns.In this paper, our main contribution is a new quasi-conformal parameterization method, based on a least-squares approximation of the Cauchy-Riemann equations. The so-defined objective function minimizes angle deformations, and we prove the following properties: the minimum is unique, independent of a similarity in texture space, independent of the resolution of the mesh and cannot generate triangle flips. The function is numerically well behaved and can therefore be very efficiently minimized. Our approach is robust, and can parameterize large charts with complex borders.We also introduce segmentation methods to decompose the model into charts with natural shapes, and a new packing algorithm to gather them in texture space. We demonstrate our approach applied to paint both scanned and modeled data sets.

Near-optimal parameterization of the intersection of quadrics

In this paper, we present the first exact, robust and practical method for computing an explicit representation of the intersection of two arbitrary quadrics whose coefficients are rational. Combining results from the theory of quadratic forms, linear algebra and number theory, we show how to obtain parametric intersection curves that are near-optimal in the number and depth of radicals involved.

A Computational Geometric Approach to Visual Hulls

Recognizing 3D objects from their 2D silhouettes is a popular topic in computer vision. Object reconstruction can be performed using the volume intersection approach. The visual hull of an object is the best approximation of an object that can be obtained by volume intersection. From the point of view of recognition from silhouettes, the visual hull can not be distinguished from the original object. In this paper, we present efficient algorithms for computing visual hulls. We start with the case of planar figures (polygons and curved objects) and base our approach on an efficient algorithm for computing the visibility graph of planar figures. We present and tackle many topics related to the query of visual hulls and to the recognition of objects equal to their visual hulls. We then move on to the 3-dimensional case and give a flavor of how it may be approached.

Intersecting quadrics: an efficient and exact implementation

We present the first complete, exact and efficient C++ implementation of a method for parameterizing the intersection of two implicit quadrics with integer coefficients of arbitrary size. It is based on the near-optimal algorithm recently introduced by Dupont et al., [2]. Unlike existing implementations, it correctly identifies and parameterizes all the connected components of the intersection in all cases, returning parameterizations with rational functions whenever such parameterizations exist. In addition, the coefficient fields of the parameterizations are either minimal or involve one possibly unneeded square root. We prove upper bounds on the size of the coefficients of the output parameterization and compare these bounds to observed values. We give other experimental results and present some examples.

The Expected Number of 3D Visibility Events Is Linear

In this paper, we show that, amongst

Line Transversals to Disjoint Balls

n

Complete, exact and efficient implementation for computing the adjacency graph of an arrangement of quadrics

uniformly distributed unit balls in

ON THE EXPECTED SIZE OF THE 2D VISIBILITY COMPLEX

\mathbb{R}^3

Line transversals to disjoint balls

, the expected number of maximal nonoccluded line segments tangent to four balls is linear. Using our techniques we show a linear bound on the expected size of the visibility complex, a data structure encoding the visibility information of a scene, providing evidence that the storage requirement for this data structure is not necessarily prohibitive. These results significantly improve the best previously known bounds of

Helly-Type Theorems for Line Transversals to Disjoint Unit Balls

O(n^{8/3})