Logical/Linear Operators in Early Vision.

Abstract

We propose a research program to investigate new techniques for computing image descriptions in early vision. The program will incorporate computational, physiological and psychophysical aspects. Research will be focused on the certain non-linearities that we have discovered can play an important role in improving the robustness of edge, line, and curve detection. These non-linearities are totally different from those previously studied in early vision. We embody them in what we call logical/linear operators, because these operators exhibit both logical (i.e., they form an appropriate Boolean algebra) and linear properties. We submit that these operators can lead to both more robust computer vision, and to new insights into biological vision.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 14, 1996
Accession Number
ADA316134

Entities

People

  • Steven W. Zucker

Organizations

  • McGill University

Tags

DTIC Thesaurus Topics

  • Boolean Algebra
  • Computer Vision
  • Computers
  • Detection
  • Linearity

Fields of Study

  • Mathematics

Readers

  • Computer Vision.
  • Graph Algorithms and Convex Optimization.
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms