Graph/Knot Theoretical Analysis and Generation for Impossible Figures

Kento Tarui, Fangyan Dong, Yutaka Hatakeyama, Kaoru Hirota

Research output: Contribution to journalArticlepeer-review

Abstract

An algorithm to represent impossible multibar figures and their subclass of torus figures is proposed based on graph and knot theory. A multibar type graph, which is an abstract concept of multibar figures, is defined by the junction graph that represents the connections of the lines. It is shown that the junction graph is able to characterize multibar figures where this characterization is realized according to the type of the multibar type graph. An automatic drawing system of torus figures is also presented by analyzing junction graphs that construct the shapes of corners of torus figures. The proposed method aims a basic tool for experiments in visual psychology and possible/impossible figures generation.

Original languageEnglish
Pages (from-to)1262-1273
Number of pages12
JournalJournal of Advanced Computational Intelligence and Intelligent Informatics
Volume11
Issue number10
DOIs
Publication statusPublished - Dec 2007
Externally publishedYes

Keywords

  • computer vision
  • graph theory
  • impossible figure
  • knot theory
  • visual psychology

Fingerprint

Dive into the research topics of 'Graph/Knot Theoretical Analysis and Generation for Impossible Figures'. Together they form a unique fingerprint.

Cite this