6533b857fe1ef96bd12b4e26

RESEARCH PRODUCT

Single-output color pattern recognition using a fractional correlator

Zeev ZalevskyGal ShabtayDavid MendlovicJavier GarciaIdo Raveh

subject

Physicsbusiness.industryColor imageGeneral EngineeringPattern recognitionComputer-generated holographyAtomic and Molecular Physics and OpticsSpectral colorsymbols.namesakeOpticsFourier transformOptical correlatorPattern recognition (psychology)symbolsRGB color modelArtificial intelligencebusinessOptical filter

description

A novel method for performing color image pattern recogni- tion using a fractional correlator (FC) is proposed. The input plane is illuminated with three different coherent sources of wavelengths corre- sponding to RGB (red, green, and blue) colors. The output plane pro- vides a single output peak, which is a result of an incoherent addition between the three correlations obtained per each color. By using the fractional correlator, which is a partially space variant correlator, we achieve space-variance-controlled color pattern recognition. The use of the three-color illumination can drastically increase the discrimination ability of the suggested correlator. © 1997 Society of Photo-Optical Instrumenta- tion Engineers. (S0091-3286(97)02808-0)

https://doi.org/10.1117/1.601428