Automatic Contour Detection by Encoding Knowledge into Active Contour Models

  • Authors:
  • Olivier Gérard;Shérif Makram-Ebeid

  • Affiliations:
  • -;-

  • Venue:
  • WACV '98 Proceedings of the 4th IEEE Workshop on Applications of Computer Vision (WACV'98)
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

An original method for an automatic detection of contoursin difficult images is proposed. This method is basedon a tight cooperation between a multi-resolution neuralnetwork and a hidden Markov model-enhanced dynamicprogramming procedure. This new method is able to overcomethe three major drawbacks of the "standard" activecontours: initialization dependancy, exclusive use of localinformation and occlusion sensitivity. The driving idea isto introduce high-order a priori information in each step ofthe system. An application to the automatic detection of theleft ventricle in digital X-ray images is proposed.