Optical pattern recognition

We discuss how one optical processor (a correlator) can be used for all levels of scene analysis (low, medium, and high-level computer vision). This is achieved by the use of different filter functions for the different levels of a hierarchical inference system. New optical processor filter research that allows such flexibility is advanced and examples of each of these filters are provided. For large class problems, feature extracted from each region of interest in a scene are fed to a neural net processor which performs recognition. New algorithms for optical neural net classifiers are also required. We conclude that present hardware optical correlator advances can significantly benefit from such processing.