Abstract
The usefulness of applying such concepts as signal-to-noise ratio to the quantitative measurement of grainy photographic images has been limited by doubts as to how this ratio depended on the measuring apparatus, and, in particular, as to how it could be related to subjective measurements. Information theory now makes it possible to treat such subjective measures in terms of objective procedures which (in the sense that they extract all the relevant information content of the image) are ideal and cannot be bettered by a human observer, but to which a skilled observer closely approximates. Examples are given illustrating how such procedures could be applied in photography. It is argued that the properties of grain can be described practically completely by the autocorrelation function of the transmission of the developed emulsion. An expression is derived giving, in terms of the autocorrelation function, the mean square fluctuation in transmission for any scanning aperture. This derivation makes no assumption as to the particular nature of grain. A rapid method of measuring the autocorrelation function is proposed.
In terms of these concepts, some discussion is given of the “root area” law of granularity, parametric representation of grain, and the relation of image sharpness to grain structure.
© 1953 Optical Society of America
Full Article | PDF ArticleMore Like This
Fred H. Perrin and J. H. Altman
J. Opt. Soc. Am. 43(9) 780-790 (1953)
Glenn A. Fry
J. Opt. Soc. Am. 53(3) 361-367 (1963)
J. H. Webb
J. Opt. Soc. Am. 38(4) 312-323 (1948)