A 3D shape descriptor based on spectral analysis of medial axis

3D shape completion is important to enable machines to perceive the complete geometry of objects from partial observations. To address this problem, view-based methods have been presented. These methods represent shapes as multiple depth images, which can be back-projected to yield corresponding 3D point clouds, and they perform shape completion by learning to complete each depth image using neural networks. While view-based methods lead to state-of-the-art results, they currently do not enforce geometric consistency among the completed views during the inference stage. To resolve this issue, we propose a multi-view consistent inference technique for 3D shape completion, which we express as an energy minimization problem including a data term and a regularization term. We formulate the regularization term as a consistency loss that encourages geometric consistency among multiple views, while the data term guarantees that the optimized views do not drift away too much from a learned shape descriptor. Experimental results demonstrate that our method completes shapes more accurately than previous techniques.

Download Full-text

Application of the curvature scale space transform to the representation of three-dimensional models

Image Processing & Communications ◽

10.2478/v10248-012-0087-y ◽

2013 ◽

Vol 18 (1) ◽

pp. 23-29

Author(s):

Dariusz Frejlichowski

Keyword(s):

Computer Aided Design ◽

Shape Representation ◽

Three Dimensional ◽

Shape Descriptor ◽

3D Models ◽

Scale Space ◽

3D Shape ◽

Three Dimensional Objects ◽

Curvature Scale Space ◽

Planar Shapes

Abstract In this paper an algorithm for the representation of 3D models is described and experimentally evaluated. Three-dimensional objects are becoming very popular recently and they are processed in various ways - analysed, retrieved, recognised, and so on. Moreover, they are employed in various aplications, such as virtual reality, entertainment, Internet, Computer Aided Design, or even in biometrics or medical imaging. That is why the development of appropriate algorithms for the representation of 3D objects is so important recently. These algorithms - so called 3D shape descriptors - are assumed to be invariant to particular transformations and deformations. One of the possible approaches is based on the projections of a 3D object into planar shapes and representation of them using a 2D shape descriptor. An algorithm realising this idea is described in this paper. Its first stage is based on the rendering of 20 2D projections, from various points of view. Later, the obtained projections are stored in a form of bitmaps and the Curvature Scale Space algorithm is applied for the description of the planar shapes extracted from them. The proposed approach is experimentally compared with several other 3D shape representation methods.

Download Full-text

A 3D Shape Descriptor Based on Depth Complexity and Thickness Histograms

2015 28th SIBGRAPI Conference on Graphics, Patterns and Images ◽

10.1109/sibgrapi.2015.51 ◽

2015 ◽

Cited By ~ 2

Author(s):

Wagner Schmitt ◽

Jose L. Sotomayor ◽

Alexandru Telea ◽

Claudio T. Silva ◽

Joao L.D. Comba

Keyword(s):

Shape Descriptor ◽

3D Shape

Download Full-text