Adaptive Filtering of Distributed Data Based on Modeling the Perception Mechanisms of Living Sensory Systems

The article is devoted to the results of the adaptive filtering synthesis in the frames of neuromorphic model inspired by biological sensory system. To adequately model the perception mechanisms the analysis and synthesis are carried out on the base of the most realistic representation of input data—external stimuli—in the form of the stream of receptor registration events. Statistical model of that stream is chosen in the form of Poisson two-dimensional point processes. On this basis, a statistical description of the input data in the form of a sampling representation is proposed. The model of neural encoding of input data is considered in the article within the framework of the receptive fields concept. A number of well-known neuro-mechanisms are implemented in the encoding model, including in particular, center/surround inhibition. Decoding issues are considered in the context of stimuli spatial contrasts restoration, which partially models the responses of so-called simple cells in the visual cortex. It is shown that the model of coupled ON-OFF-decoding admits the sharp image details restoring in the form of local edges. At the end of the article, to justify the adequacy of the synthesized encoding procedure—adaptive filtering—we demonstrate an example of image edge directed interpolation and discus the results in the terms of perceptual quality.