Learning Shape Placements by Example

Paul Guerrero, Stefan Jeschke, Michael Wimmer, Peter Wonka
Learning Shape Placements by Example
ACM Transactions on Graphics, 34(4):108:1-108:13, August 2015. [additional_material] [paper] [video]

Information

Abstract

We present a method to learn and propagate shape placements in 2D polygonal scenes from a few examples provided by a user. The placement of a shape is modeled as an oriented bounding box. Simple geometric relationships between this bounding box and nearby scene polygons define a feature set for the placement. The feature sets of all example placements are then used to learn a probabilistic model over all possible placements and scenes. With this model we can generate a new set of placements with similar geometric relationships in any given scene. We introduce extensions that enable propagation and generation of shapes in 3D scenes, as well as the application of a learned modeling session to large scenes without additional user interaction. These concepts allow us to generate complex scenes with thousands of objects with relatively little user interaction.

Additional Files and Images

Additional images and videos

Additional files

additional_material: additional material additional_material: additional material
paper_lowres: pre-print version of the paper in low resolution paper_lowres: pre-print version of the paper in low resolution
paper: pre-print version of the paper paper: pre-print version of the paper

Weblinks

BibTeX

@article{guerrero-2015-lsp,
  title =      "Learning Shape Placements by Example",
  author =     "Paul Guerrero and Stefan Jeschke and Michael Wimmer and
               Peter Wonka",
  year =       "2015",
  abstract =   "We present a method to learn and propagate shape placements
               in 2D polygonal scenes from a few examples provided by a
               user. The placement of a shape is modeled as an oriented
               bounding box. Simple geometric relationships between this
               bounding box and nearby scene polygons define a feature set
               for the placement. The feature sets of all example
               placements are then used to learn a probabilistic model over
               all possible placements and scenes. With this model we can
               generate a new set of placements with similar geometric
               relationships in any given scene. We introduce extensions
               that enable propagation and generation of shapes in 3D
               scenes, as well as the application of a learned modeling
               session to large scenes without additional user interaction.
               These concepts allow us to generate complex scenes with
               thousands of objects with relatively little user
               interaction.",
  month =      aug,
  issn =       "0730-0301",
  journal =    "ACM Transactions on Graphics",
  number =     "4",
  volume =     "34",
  pages =      "108:1--108:13",
  keywords =   "modeling by example, complex model generation",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2015/guerrero-2015-lsp/",
}