Correlation-based evaluation of visual performance to reduce the statistical error of visual acuity.

Ophthalmologists evaluate visual acuity tests by the number of correctly recognized optotypes (usually letters) in the different lines of an eye chart. This probability-based scoring results in significant statistical error that can only be decreased by the time-consuming analysis of a larger number of optotypes. In this paper, we present a new, more precise correlation-based scoring method that takes the degree of misidentification into consideration too, rather than the mere fact of it. According to our experimental results, this new method decreases the uncertainty error by 28% if using the same number of optotypes at a given letter size or requires half the optotype number to produce the same error as that of probability-based scoring.