Visual acuity: Information theory, retinal image structure and resolution thresholds

The traditional view of the resolving power of the eye as based on the Rayleigh Criterion needs extension in view of insight gained from information theory. As first pointed out by Toraldo di Francia, the image of a double line is subtly different from that of a single line even when there is no central dip. Given an image, the uncertainty concerning the generating object can be expressed numerically in bits. With the available data on the optical performance and anatomical structure of the normal human eye, details of the distribution of light and of receptor activation have been characterized for foveal two-line resolution. They allow a comparison with the actual performance of the human observer and some conclusions concerning the strategy employed in resolution decisions as well as a more comprehensive analysis of the processes involved in the measurement of visual acuity. The discussion is extended to the case of laser-illuminated targets where the difference between image formation in coherent and incoherent light becomes relevant.