The brightness of noisy textures

Abstract

It is well known that the brightness of a surface depends on a variety of factors. Luminance is one of the main determinants of this brightness and its articulation in the image may produce changes in the perception. We investigate how a patch, which was textured with luminance noise, is perceived in comparison with an homogeneous one. Our hypothesis was that the two patches must be perceived as equally bright when their mean luminance were the same. To test this hypothesis, we performed two experiments. First, we estimated by mean of flicker photometry the brightness equality between textured and homogeneous patches. Secondly, we used a two-alternatives-unforced-choice paradigm to measure the brightness PSE between textured and homogeneous patches. The first experiment showed that subjects found the minimum flicker situation when textured and homogeneous patches had the same mean luminance, which indicates that the visual system is effectively sensing mean luminance. However, the results of the second experiment show a systematic perceptual bias indicating that the textured patch was perceived darker than the homogeneous patch. This result suggests that textures modify the brightness of surfaces perhaps, because its interaction with the background is different than when the surface is homogeneous.