Temporal Correlation Theory and Modeling the Segmentation of the Visual Information in the Brain (a review)
Institute of Mathematical Problems of Biology, Russian Academy of Sciences, Pushchino, Moscow Region, 142292, Russia
Abstract. A review of neural network models for image segmentation is presented. The models are based on dynamic processes, oscillations, and synchronization of neural activity. Neurobiological principles of the models, the mechanisms of their functioning, and the results of simulations are described. The performance of the models and the perspectives of improving the segmentation quality are discussed.
Key words: image segmentation, oscillatory neural networks, synchronization, feature binding, memory, attention.