Silicon models of lateral inhibition

Abstract

Lateral inhibition is characterized in two CMOS VLSI-based sensory arrays with different pixel geometries and inhibitory algorithms. In these arrays, each sensory pixel inhibits, and in turn, is inhibited by each of its immediate neighbors. Test results indicate that both circuits are able to enhance the contrast of arbitrary two-dimensional images, and do so in a manner that is dynamically stable. Achieving this enhancement on sensory images immediately and simultaneously, this operation offers performance at a level not obtainable by software methods. As such, it is well suited for machine vision systems that utilize parallel architectures.<>