1. Science
  2. Publications
  3. Scientific Works of Kharkiv National Air Force University
  4. 1(19)'2009
  5. Machine learning methodological concepts for irregular object image segmentation

Machine learning methodological concepts for irregular object image segmentation

К.S. Smelyakov
Annotations languages:

In order to increase a stability of irregular object image segmentation in condition of uncertainty for autonomous vision systems the machine learning models are proposed which are based on forecasting functions and image templates for the considered objects and scenes.
Keywords: segmentation, teaching, self-training, prognosis, standard