Premium
Discrete‐time dynamic image segmentation based on oscillations by destabilizing a fixed point
Author(s) -
Fujimoto Ken'ichi,
Kobayashi Mio,
Yoshinaga Tetsuya
Publication year - 2011
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.20683
Subject(s) - segmentation , computer science , fixed point , image (mathematics) , artificial intelligence , computer vision , image segmentation , control theory (sociology) , mathematics , control (management) , mathematical analysis
We have developed a discrete‐time dynamical system consisting of a global inhibitor and chaotic neurons that can generate oscillatory responses. We have also found that our system can work as a dynamic image‐segmentation system utilizing oscillatory responses of chaotic neurons. Dynamic image segmentation is to severally extract isolated image regions from a static image and is to exhibit segmented images in time series according to oscillatory responses of chaotic neurons. At certain parameter values, chaotic neurons can generate adequate oscillatory responses for dynamic image segmentation. However, they generate non‐oscillatory responses for certain initial values due to the coexistence of a stable fixed point corresponding to non‐oscillatory responses. Their appearances indicate that our system does not work as a dynamic image‐segmentation system. In this study, we designed a destabilizer for a stable fixed point to prevent its appearance. We also demonstrated that our system with a designed destabilizer worked well. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.