Human pattern recognition as a holographic process

Abstract

A novel analysis of the human pattern recognition process as a holographic associative memory is presented. The response of the two recognition systems to partial input images is investigated. Simulations based on holographic cross correlation show that a power law describes the decrease in recognition strength as the number of pixels in an input object is reduced. Human recognition response is studied by conducting a survey among 400 respondents. Power law behavior in the recognition success rate is observed from 12.5% to 62.5% image content, with an exponent of 1.65.