Technical Papers

Points & Reconstruction

Monday, 11 August 3:45 PM - 5:15 PM | Vancouver Convention Centre, East Building, Ballroom A Session Chair: Michael Kazhdan, Johns Hopkins University

Point Morphology

A new approach to the morphological analysis of point clouds, based on a new structuring element model, a robust projective formulation of the operators, and a specific adaptive sampler. The method is applied to geometry and topology manipulation, including medial axis sampling, hysteresis shape filtration, and topological simplification.

Stéphane Calderon
Télécom ParisTech

Tamy Boubekeur
Télécom ParisTech

Floating Scale Surface Reconstruction

Introducing a novel, virtually parameter-free method for surface reconstruction from oriented, scale-enabled point samples. The approach constructs an implicit function as the sum of compactly supported basis functions. The final surface is extracted as the zero-level set of the implicit function.

Simon Fuhrmann
Technische Universität Darmstadt

Michael Goesele
Technische Universität Darmstadt

Continuous Projection for Fast L1 Reconstruction

Continuous locally optimal projection (CLOP) is a novel analytic projection operator that allows for fast L1 reconstruction of dynamic point sets at interactive frame rates. It uses a continuous reformulation of the discrete weighted locally optimal projection (WLOP) operator, which it outperforms by up to an order of magnitude.

Reinhold Preiner
Technische Universität Wien

Oliver Mattausch
Universität Zürich

Murat Arikan
Technische Universität Wien

Renato Pajarola
Universität Zürich

Michael Wimmer
Technische Universität Wien

Flower Modeling Via X-Ray Computed Tomography

This paper presents a novel flower-modeling technique that utilizes X-ray computed tomography and real-world flowers. The method approximates a flower by key primitives, a shaft, and a sheet, and semi-automatically fits them to flower CT volume by using new active curve and surface models.

Takashi Ijiri

Shin Yoshizawa

Hideo Yokota

Takeo Igarashi
The University of Tokyo

k-d Darts: Sampling by k-Dimensional Flat Searches

This paper extends the concept of point sampling to hyperplane sampling: lines, planes, etc. Samples are axis-aligned yet unbiased. The paper demonstrates high-dimensional applications in disk packing, rendering, and failure analysis.

Mohamed Ebeida
Sandia National Laboratories