VISUAL SIGNAL DETECTABILITY WITH 2 NOISE COMPONENTS - ANOMALOUS MASKING EFFECTS

We measured human observers' detectability of aperiodic signals in noi se with two components (white and low-pass Gaussian). The white-noise component ensured that the signal detection task was always noise limi ted rather than contrast limited (i.e., image noise was always much la rger than observer internal noise), The low-pass component can be cons idered to be a statistically defined background. Contrast threshold el evation was not linearly related to the rms background contrast. Our r esults gave power-law exponents near 0.6, similar to that found for de terministic masking. The Fisher-Hotelling linear discriminant model as sessed by Rolland and Barrett [J. Opt. Soc. Am. A 9, 649 (1992)] and t he modified nonprewhitening matched filter model suggested by Burgess [J. Opt. Sec. Am. A 11, 1237 (1994)] for describing signal detection i n statistically defined backgrounds did not fit our more precise data. We show that it is not possible to find any nonprewhitening model tha t can fit our data. We investigated modified Fisher-Hotelling models b y using spatial-frequency channels, as suggested by Myers and Barrett [J. Opt. Sec. Am. A 4, 2447 (1987)]. Two of these models did give good fits to our data, which suggests that we may be able to do partial pr ewhitening of image noise. (C) 1997 Optical Society of America.