CNNUM stereo architecture and 3D template design techniques

Tamáa Szirányi, László Czúni

Research output: Contribution to journalArticle

1 Citation (Scopus)


A CNN Universal Machine-based structure and analogic algorithm is proposed to extract 3D spatial information from stereo images. The CNN template values are obtained in two ways: from simplex optimization and from the energy function. The method is demonstrated on densely textured as well as grey scale stereograms.

Original languageEnglish
Pages (from-to)25-41
Number of pages17
JournalInternational Journal of Circuit Theory and Applications
Issue number1
Publication statusPublished - Jan 1 1999



  • Cellular neural networks
  • Stereo depth detection

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Applied Mathematics

Cite this