A neural-like layer for processing dynamic images

Summary form only given. A neural-like layer with dynamic mechanisms employed in the processing of continually changing visual images is discussed. The information embedded in the visual image is assumed to be dynamic in both space and time. The processing of information by dynamic mechanisms is imperative because a conventional static vision system fails to achieve the true adaptability and robustness possessed by even the simplest biological system. Visual information is processed over a hierarchical network of planar regions of nerve cells called neural layers. The neural layer is a collective assembly of numerous interconnected neural subunits. Each neural subunit consists of several localized dynamic neural sets. An individual neural set is modeled as an underdamped second-order system cascaded with a static nonlinear transformation function representing the dynamics of parallel noninteractive biological neuron assemblies.<>