Pattern processing and symbolic processing are the major methods used in an image understanding task. Conventionally, they are often implemented and handled as independent systems. However, the handling of real images requires a method that incorporates the characteristics of both. But the designing of a method that enables interaction between symbolic processing and pattern processing is not an easy task. In this paper, we propose a method for symbol-pattern mutual transformation, which, through symbolization of the connection knowledge acquired by the Selective Attention Model, lays the foundation for symbol-pattern integration. We demonstrate the model's effectiveness by applying it to an object segmentation problem.
Recent studies have revealed that in vivo cortical neurons show spontaneous ‘UP-DOWN’ transitions between the two subthreshold levels of the membrane potentials, i.e., ‘UP’ state and ‘DOWN’ state. The neural mechanism of generating those spontaneous state transitions, however, remains unclear. Recent electrophysiological studies suggested that those state transitions may occur through activation of a hyperpolarization-activated cation current (H-current, Ih) by inhibitory synaptic inputs (Cossart et al., 2002). To show that the spontaneous state transitions can be generated by a network-based mechanism, we study learning processes in a computational model of cortical networks. We now found that the spontaneous ‘UP-DOWN’ transitions similar to those exhibited by in vivo neurons can be self-organized through spike-timing-dependent plasticity in a network of inhibitory neurons and excitatory neurons expressing the H-current.
The stereo vision system derives a 3D representation of the visual world from a pair of 2D representations of retinal images. For this well-defined computational task, stereo vision serves as a model system for investigations of the neural mechanisms of visual perception. Computational ideas merge with neurophysiological data of single neuron responses in the monkey visual cortex. Area V1 performs initial filter-like processing, but this alone is not sufficient for all aspects of stereo vision. Several higher visual areas play important roles. Recent findings suggest that areas along the dorsal and ventral pathways have complementary functions in stereopsis.