A pattern recognition system based on models of aspects of the human visual system

摘要

The natural world contains many examples of successful vision systems embodying important design principles and remarkable optimalities. There is much to be learned by studying and modelling natural vision systems. The algorithm discussed illustrates how aspects of even the present partial understanding of human vision can be applied to pattern representation and recognition. The authors review the log-polar model for the sampling structure of the human retina, then consider the application of the model in an algorithm for scale and rotation invariant recognition of patterns.