Relation-Based Parametrization and Exploration of Shape Collections

Kurt Leimer, Lukas Gersthofer, Michael Wimmer, Przemyslaw Musialski
Relation-Based Parametrization and Exploration of Shape Collections
Computers & Graphics, 67():127-137, October 2017. [preprint] [video]

Information

Abstract

With online repositories for 3D models like 3D Warehouse becoming more prevalent and growing ever larger, new possibilities have emerged for both experienced and inexperienced users. These large collections of shapes can provide inspiration for designers or make it possible to synthesize new shapes by combining different parts from already existing shapes, which can be both easy to learn and a fast way of creating new shapes. But exploring large shape collections or searching for particular kinds of shapes can be difficult and time-consuming tasks as well, especially considering that online repositories are often disorganized. In our work, we propose a relation-based way to parametrize shape collections, allowing the user to explore the entire set of shapes based on the variability of spatial arrangements between pairs of parts. The way in which shapes differ from each other is captured automatically, resulting in a small number of exploration parameters. Furthermore, a copy-and-paste system for parts allows the user to change the structure of a shape, making it possible to explore the entire collection from any initial shape.

Additional Files and Images

Additional images and videos

Additional files

Weblinks

BibTeX

@article{leimer_2017_rbpesc,
  title =      "Relation-Based Parametrization and Exploration of Shape
               Collections",
  author =     "Kurt Leimer and Lukas Gersthofer and Michael Wimmer and
               Przemyslaw Musialski",
  year =       "2017",
  abstract =   "With online repositories for 3D models like 3D Warehouse
               becoming more prevalent and growing ever larger, new
               possibilities have emerged for both experienced and
               inexperienced users. These large collections of shapes can
               provide inspiration for designers or make it possible to
               synthesize new shapes by combining different parts from
               already existing shapes, which can be both easy to learn and
               a fast way of creating new shapes. But exploring large shape
               collections or searching for particular kinds of shapes can
               be difficult and time-consuming tasks as well, especially
               considering that online repositories are often disorganized.
               In our work, we propose a relation-based way to parametrize
               shape collections, allowing the user to explore the entire
               set of shapes based on the variability of spatial
               arrangements between pairs of parts. The way in which shapes
               differ from each other is captured automatically, resulting
               in a small number of exploration parameters. Furthermore, a
               copy-and-paste system for parts allows the user to change
               the structure of a shape, making it possible to explore the
               entire collection from any initial shape.",
  month =      oct,
  issn =       "0097-8493",
  journal =    "Computers & Graphics",
  volume =     "67",
  pages =      "127--137",
  keywords =   "3d database exploration, Model variability, Shape analysis,
               Shape collections",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2017/leimer_2017_rbpesc/",
}