PHOTOGRAPHIC GRAIN NOISE SUPPRESSION BY DENSITY QUANTIZATION: ITS INFLUENCE ON IMAGE QUALITY

Abstract

A technique is described for suppressing unwanted grain noise in scanned and digitized photographic images. The technique employs the rms granularity statistics of the film to divide, or quantize, the density range of the digital image into distinguishable ranges of density. For Eastman Kodak type 3414 film, the rms granularity statistics determined indicate constant Selwyn granularity for the range of scanning spot diameters from approximately 16 to 4 micrometers, and this result agrees with Kodak data for a 48 micrometer diameter spot. For spots from 16 to 4 micrometers diameter, the rms granularity is also determined to be a function of the square root of diffuse density. The number of distinguishable density levels, required to sup- press the grain noise of the digitally processed image, is found to be directly related to the scanning spot diameter. Finally, evaluation of the resulting digitally processed images indicates that subjective image quality is directly related to the spatial resolution of the image. That is, with the grain noise suppressed, subjective quality is improved by scanning the original image with a smaller diameter spot.