Natural Image Statistics

Binocular rivalry stimuli are common but rivalry is not

by Robert O'Shea

O’Shea, R. P. (2011). Binocular rivalry stimuli are common but rivalry is not. Frontiers in Human Neuroscience, 5(148), 1-2. doi: 10.3389/fnhum.2011.00148

Recently Arnold (2011) asked “Why is binocular rivalry uncommon?”. He answered in an entertainingly written,... more Recently Arnold (2011) asked “Why is binocular rivalry uncommon?”. He answered in an entertainingly written, provocative article, for which I thank and congratulate him. However, I will argue that Arnold’s answer falls short in two respects and his assumption that rivalry is uncommon is correct for two reasons other than the one he discusses.

POSTER at AVA/BMVA 2012 in Cambridge: Image discomfort: its biological basis and utility - a review

by Sarah M Haigh

Arnold J Wilkins and Sarah M Haigh
A review of published and in prep work

Observers show consistency in the images they judge to be uncomfortable. The discomfort can be predicted from the... more Observers show consistency in the images they judge to be uncomfortable. The discomfort can be predicted from the statistics of the image, and this applies to images ranging from photographs of everyday scenes to geometric arrays, coloured or in motion. Images are generally rated as uncomfortable to view if (1) the luminance contrast energy has a power spectrum that departs from 1/f, or (2) the CIE UCS chromaticities are widely separated, or (3) the contours flicker at about 20Hz. For the same variety of images, those that are uncomfortable result in a large cortical haemodynamic response, measured either with fMRI or near infrared spectroscopy. Individual differences in the size of the haemodynamic response appear to reflect a cortical hyperexcitability, common in migraine, and providing clues as to the mechanisms of discomfort. The biological utility of the discomfort can be understood not only as homeostatic, but also as an evolutionary adaptation because images of venomous animals possess the properties of uncomfortable images and are indeed uncomfortable to view. There are far-reaching implications for the design of our visual environment because much of our environment has characteristics that are uncomfortable.

Spectral in-homogeneity provides information for figure ground organization

by Sudarshan Ramenahalli

Neuroscience 2011, Washington DC, USA

Localized information is necessary for scene categorization, including the Natural/Man-made distinction

by Lester Loschky

What information do people use to categorize scenes? Computational scene classification models have proposed that... more What information do people use to categorize scenes? Computational scene classification models have proposed that unlocalized amplitude information, the distribution of spatial frequencies and orientations, is useful for categorizing scenes. Previous research has provided conflicting results regarding this claim. Our previous research (Loschky et al., 2007) has shown that randomly localizing amplitude information (i.e., randomizing phase) greatly disrupts scene categorization at the basic level. Conversely, studies suggesting the usefulness of unlocalized amplitude information have used binary distinctions, e.g., Natural/Man-made. We hypothesized that unlocalized amplitude information contributes more to the Natural/Man-made distinction than basic level distinctions. Using an established set of images and categories, we varied phase randomization and measured participants’ ability to distinguish Natural versus Man-made scenes or scenes at the basic level. Results showed that eliminating localized information by phase randomization disrupted scene classification even for the Natural/Man-made distinction, demonstrating that amplitude localization is necessary for scene categorization.

The importance of information localization in scene gist recognition.

by Lester Loschky

People can recognize the meaning or gist of a scene from a single glance, and a few recent studies have begun to... more People can recognize the meaning or gist of a scene from a single glance, and a few recent studies have begun to examine the sorts of information that contribute to scene gist recognition. The authors of the present study used visual masking coupled with image manipulations (randomizing phase while maintaining the Fourier amplitude spectrum; random image structure evolution [RISE]; J. Sadr & P. Sinha, 2004) to explore whether and when unlocalized Fourier amplitude information contributes to gist perception. In 4 experiments, the authors found that differences between scene categories in the Fourier amplitude spectrum are insufficient for gist recognition or gist masking. Whereas the global 1/f spatial frequency amplitude spectra of scenes plays a role in gist masking, local phase information is necessary for gist recognition and for the strongest gist masking. Moreover, the ability to recognize the gist of a target image was influenced by mask recognizability, suggesting that conceptual masking occurs even at the earliest stages of scene processing.

The role of higher order image statistics in masking scene gist recognition

by Lester Loschky

In the present article, we investigated whether higher order image statistics, which are known to be carried by the... more In the present article, we investigated whether higher order image statistics, which are known to be carried by the Fourier phase spectrum, are sufficient to affect scene gist recognition. In Experiment 1, we compared the scene gist masking strength of four masking image types that varied in their degrees of second- and higher order relationships: normal scene images, scene textures, phase-randomized scene images, and white noise. Masking effects were the largest for masking images that possessed significant higher order image statistics (scene images and scene textures) as compared with masking images that did not (phase-randomized scenes and white noise), with scene image masks yielding the largest masking effects. In a control study, we eliminated all differences in the second-order statistics of the masks, while maintaining differences in their higher order statistics by comparing masking by scene textures rather than by their phase-randomized versions, and showed that the former produced significantly stronger gist masking. Experiments 2 and 3 were designed to test whether conceptual masking could account for the differences in the strength of the scene texture and phase-randomized masks used in Experiment 1, and revealed that the recognizability of scene texture masks explained just 1% of their masking variance. Together, the results suggest that (1) masks containing the higher order statistical structure of scenes are more effective at masking scene gist processing than are masks lacking such structure, and (2) much of the disruption of scene gist recognition that one might be tempted to attribute to conceptual masking is due to spatial masking.

Mixtures of conditional Gaussian scale mixtures applied to multiscale image representations

by Lucas Theis

In All Likelihood, Deep Belief Is Not Enough

by Lucas Theis

Natural Visualizations

by Steve Haroz

Published in EuroVis 2006

This paper demonstrates the prevalence of a shared characteristic between visualizations and images of nature. We have... more This paper demonstrates the prevalence of a shared characteristic between visualizations and images of nature. We have analyzed visualization competitions and user studies of visualizations and found that the more preferred, better performing visualizations exhibit more natural characteristics. Due to our brain being wired to perceive natural images [SO01], testing a visualization for properties similar to those of natural images can help show how well our brain is capable of absorbing the data. In turn, a metric that finds a visualization’s similarity to a natural image may help determine the effectiveness of that visualization. We have found that the results of comparing the sizes and distribution of the objects in a visualization with those of natural standards strongly correlate to one’s preference of that visualization.