Another look at Cellular Neural Networks

The CNN architecture, introduced in 1988, mimicked the retina's functionality. Our understanding of how the sensory nervous system works, including vision, has significantly improved. At the same time, both integrated circuit design technology and signal processing have made extraordinary progress. While the latter have enabled numerous applications, traditional paradigms of sensing and real time processing are reaching a breakpoint caused by a deluge of sensory data requiring excessive processing power to extract the underlying information that leads to desirable action. Fundamental breakthroughs are required to overcome multiple obstacles presently limiting the pace of progress. In this contribution it is advocated that renewing the CNN architecture, enabled by today's technologies, and inspired by clearer knowledge of the visual system, is a better way to resolve data acquisition and processing paradigm's bottlenecks in a variety of emerging industrial applications.