论文信息 - Locally coupled chaotic oscillator network for scene segmentation

Locally coupled chaotic oscillator network for scene segmentation

A chaotic oscillatory correlation network for scene segmentation is presented. It is a two-dimensional array with locally coupled chaotic elements. It offers a mechanism to escape the synchrony-desynchrony dilemma. As a result, this model has unbounded capacity of segmentation. Chaotic dynamics in the periodically driven Wilson-Cowan oscillator is analyzed by calculating the Lyapunov exponents. Computer simulations confirm the theoretical prediction.

[1] W. Singer,et al. Oscillatory responses in cat visual cortex exhibit inter-columnar synchronization which reflects global stimulus properties , 1989, Nature.

[2] Carroll,et al. Synchronization in chaotic systems. , 1990, Physical review letters.

[3] W. Singer,et al. Interhemispheric synchronization of oscillatory neuronal responses in cat visual cortex , 1991, Science.

[4] H. Fujisaka,et al. Stability Theory of Synchronized Motion in Coupled-Oscillator Systems , 1983 .

[5] A. Wolf,et al. Determining Lyapunov exponents from a time series , 1985 .

[6] B. Baird. Nonlinear dynamics of pattern formation and pattern recognition in the rabbit olfactory bulb , 1986 .

[7] Deliang Wang,et al. Global competition and local cooperation in a network of neural oscillators , 1995 .

[8] Ch. von der Malsburg,et al. A neural cocktail-party processor , 1986, Biological Cybernetics.

[9] Kunihiko Kaneko,et al. Relevance of dynamic clustering to biological networks , 1993, chao-dyn/9311008.

[10] Hansel,et al. Synchronization and computation in a chaotic neural network. , 1992, Physical review letters.

[11] DeLiang Wang,et al. Synchronization and desynchronization in a network of locally coupled Wilson-Cowan oscillators , 1996, IEEE Trans. Neural Networks.