The Effect of Contrast Randomisation on the Discrimination of Changes in the Slopes of the Amplitude Spectra of Natural Scenes

Abstract

It has been suggested (Tadmor and Tolhurst, 1994 Vision Research 34 541-554) that the psychophysical task of discriminating changes in the slope of the amplitude spectrum of a complex image may be similar to detecting differences in the degree of blur. It has also been suggested that human observers may perform this discrimination by detecting changes in the effective contrast within single narrow spatial-frequency bands, rather than by detecting changes in the slope per se which would involve the use of contrast information across many different frequency bands. To distinguish between these two possibilities, we have developed an experiment where observers were asked to discriminate changes in the spectral slope while different amounts of random contrast variation were introduced, with the purpose of disrupting their performance. This disruptive effect was designed to be particularly manifest if the observer really was performing a single-frequency-band contrast discrimination but to be unnoticeable if the observer was discriminating the change of slope per se. Our results imply that the observers do not usually detect changes in contrast in just one narrow spatial-frequency band when they discriminate changes in the slope of the amplitude spectrum. Rather, they must compare contrast between bands or, at least, they use contrast information from more than one band. However, for edge-enhanced (whitened) pictures, there is some evidence to suggest that observers rely on contrast changes in only a limited low-spatial-frequency band.

Evaluation of a multiscale color model for visual difference prediction

“…We estimated the observer's contrast discrimination functions for achromatic gratings indirectly by adjusting the position on the x-axis (contrast reference) and y-axis (contrast difference) of a "dipper function" template for contrast discrimination according to the observer's contrast detection thresholds measured for a grating of the same spatial frequency [Párraga and Tolhurst 2000]. In fact, the dipper template used within the model has a different form from the experimental one; the model template is adjusted so that, on solution of the model for discriminating contrast gratings, the model's output will have the same form as the measured experimental data.…”

“…We, too, have been developing a simple (low-level), physiologically plausible model of achromatic local contrast discrimination to predict human performance for discriminating between pairs of slightly different achromatic natural-scene-based images [Tadmor and Tolhurst 1994;Tadmor 1997a, 1997b;Párraga and Tolhurst 2000;Párraga et al 2005]. We compute the band-limited contrast [Peli 1990] at several spatial scales within images.…”

“…We have described a variety of psychophysical experiments, measuring thresholds for discriminating small changes in naturalistic images that we control by morphing . We decided to use a morphing technique (as opposed, to a superimposition of two images to different degrees [e.g., Tolhurst and Tadmor 2000], because it produces a set of stimuli where each one of the component pictures is an image of a plausible object (with slightly different shape, color, and texture); each morphed image still shares the natural Fourier statistics of the original ones [see Párraga and Tolhurst 2000].…”

Human visual processing of orientation in broad stimuli.

“…As will be discussed in later sections of the current paper, the exponent of content-amplitude fall-off can determine the scale at which the visual system is most sensitive as well as determine the ability of humans to discriminate between different content in different natural scenes (Knill, Field, & Kersten, 1990;Tadmor & Tolhurst, 1994;Tolhurst & Tadmor, 1997;Webster & Miyahara, 1997;Parraga, Troscianko, & Tolhurst, 2000;Parraga & Tolhurst, 2000;Tolhurst & Tadmor, 2000). In the literature devoted to measuring the distribution of content in large samples of real-world imagery (Le., determine the typical a encountered in natural scenes), the image sets generally consisted of imagery ranging from purely naturalistic content (i.e., woodlands, meadows, general shrubbery, etc.)…”

“…The task that was devised by Parraga et al (2000) involved progressively "morphing" two natural images that were very different in perceptual meaning (e.g. car and bull or the face of a man and the face of a woman), but had similar salient features.…”