ChapterPDF Available

The “Horizontal Effect”: A perceptual anisotropy in visual processing of naturalistic broadband stimuli

Authors:

Figures

Content may be subject to copyright.
A preview of the PDF is not available
... This effect is termed contrast constancy. d) Orientation: It has been observed that natural scene imagery contains a larger proportion of frequency content in vertical and horizontal orientations (i.e. the cardinal axes) relative to the oblique orientations [30], [31] (with an additional slight bias to horizontal content as noted by Hansen and Essock [32]). This is reflected in the orientational tuning of the contrast sensitivity function of the HVS. ...
... Initial studies have shown that the contrast sensitivity function obtained using simple grating thresholds is more sensitive to horizontal and vertical orientations (the so called "Oblique Effect" [33], [34]). However, more recent experiments [30], [32] using broadband natural image masking content have conversely shown that the visual system is more sensitive to oblique orientations. It is anticipated that the conventional "Oblique Effect" will be reflected in our results. ...
Article
Full-text available
Accurate estimation of the contrast sensitivity of the human visual system is crucial for perceptually based image processing in applications such as compression, fusion and denoising. Conventional Contrast Sensitivity Functions (CSFs) have been obtained using fixed sized Gabor functions. However, the basis functions of multiresolution decompositions such as wavelets often resemble Gabor functions but are of variable size and shape. Therefore to use conventional contrast sensitivity functions in such cases is not appropriate. We have therefore conducted a set of psychophysical tests in order to obtain the contrast sensitivity function for a range of multiresolution transforms: the Discrete Wavelet Transform (DWT), the Steerable Pyramid, the Dual-Tree Complex Wavelet Transform (DT-CWT) and the Curvelet Transform. These measures were obtained using contrast variation of each transforms' basis functions in a 2AFC experiment combined with an adapted version of the QUEST psychometric function method. The results enable future image processing applications that exploit these transforms such as signal fusion, super-resolution processing, denoising and motion estimation, to be perceptually optimised in a principled fashion. The results are compared to an existing vision model (HDR-VDP2) and are used to show quantitative improvements within a denoising application compared to using conventional CSF values.
... Past research studied psychophysical aspects of visual anisotropy [HHE08,OVW11] and created statistical models describing the relationship between perceived and computational textural anisotropy [Koe84,WH04,OVW11] or directionality [SPGC08]. The prediction of anisotropic highlights locations has already been studied [LKK00] and recently further extended to arbitrary geometry with interactive tangents editing [RGB*14]. ...
Article
Full-text available
Coatings are used today for products, ranging from automotive production to electronics and everyday use items. Product design is taking on an increasingly important role, where effect pigments come to the fore, offering a coated surface extra optical characteristics. Individual effect pigments have strong anisotropic, azimuthaly‐dependent behaviour, typically suppressed by a coating application process, randomly orienting pigment particles resulting in isotropic appearance. One exception is a pigment that allows control of the azimuthal orientation of flakes using a magnetic field. We investigate visual texture effects due to such an orientation in a framework allowing efficient capturing, modelling and editing of its appearance. We captured spatially‐varying BRDFs of four coatings containing magnetic effect pigments. As per‐pixel non‐linear fitting cannot preserve coating sparkle effects, we suggest a novel method of anisotropy modelling based on images shifting in an angular domain. The model can be utilized for a fast transfer of desired anisotropy to any isotropic effect coating, while preserving important spatially‐varying visual features of the original coating. The anisotropic behaviour was fitted by a parametric model allowing for editing of coating appearance. This framework allows exploration of anisotropic effect coatings and their appearance transfer to standard effect coatings in a virtual environment.
... Thomson and Foster (1997), however, used natural and phase randomized natural scenes and physically modulated the a of these images, which preserves biases in contrast across orientation even for phase scrambled images. Humans are not equally sensitive to orientation and in broadband images as they have worse sensitivity to cardinal than oblique orientations Hansen et al., 2015;Hansen, Haun, & Essock, 2008;Hansen & Essock, 2006;Haun & Essock, 2010). When measuring a discrimination thresholds with natural scenes, it is likely that any biases in sensitivity to oriented contrast will influence the ability of observers to correctly complete the a discrimination task and consequently alter tuning to a and may account for any discrepency between our findings and those of Thomson and Foster (1997). ...
Article
Full-text available
Spatial summation of luminance contrast signals has historically been psychophysically measured with stimuli isolated in spatial frequency (i.e., narrowband). Here, we revisit the study of spatial summation with noise patterns that contain the naturalistic 1/fα distribution of contrast across spatial frequency. We measured amplitude spectrum slope (α) discrimination thresholds and verified if sensitivity to α improved according to stimulus size. Discrimination thresholds did decrease with an increase in stimulus size. These data were modeled with a summation model originally designed for narrowband stimuli (i.e., single detecting channel; Baker & Meese, 2011; Meese & Baker, 2011) that we modified to include summation across multiple-differently tuned-spatial frequency channels. To fit our data, contrast gain control weights had to be inversely related to spatial frequency (1/f); thus low spatial frequencies received significantly more divisive inhibition than higher spatial frequencies, which is a similar finding to previous models of broadband contrast perception (Haun & Essock, 2010; Haun & Peli, 2013). We found summation across spatial frequency channels to occur prior to summation across space, channel summation was near linear and summation across space was nonlinear. Our analysis demonstrates that classical psychophysical models can be adapted to computationally define visual mechanisms under broadband visual input, with the adapted models offering novel insight on the integration of signals across channels and space.
... Most of the work relating to anisotropy in computer vision deals with textural information. Past research studies psychophysical aspects of visual anisotropy [Hansen et al. 2008], [Ons et al. 2011] and creates statistical models describing the relationship between perceived and computational textural anisotropy [Roli 1996], [Wenk and Houtte 2004], [Ons et al. 2011] or directionality [Shah et al. 2008]. Anisotropy has been also used as a metric for image fidelity and quality assessment [Gabarda and Cristóbal 2007]. ...
Conference Paper
Full-text available
The majority of the materials we encounter in the real-world have variable reflectance when rotated along a surface normal. This view and illumination azimuthally-variable behavior is known as visual anisotropy. Such behavior can be represented by a four-dimensional anisotropic BRDF that characterizes the anisotropic appearance of homogeneous materials. Unfortunately, most past research has been devoted to simplistic three dimensional isotropic BRDFs. In this paper, we analyze and categorize basic types of BRDF anisotropy, use a psychophysical study to assess at which conditions can isotropic appearance be used without loss of details in material appearance. To this end, we tested the human impression of material anisotropy on various shapes and under two illuminations. We conclude that subjects sensitivity to anisotropy declines with increasing complexity of 3D geometry and increasing uniformity of illumination environment. Finally, we derive and perceptually validate a computationally efficient measure of material visual anisotropy.
... However, that work has only investigated suppression as a result of narrowband 2 targets and masks and can therefore only offer a simplifi ed view of how different populations of fi lters may interact under conditions where more than a handful would be active. That is, the retinal image formed by real-world environments is known to be broadband in both spatial frequency and orientation (reviewed in Hansen et al., 2008 ), which means that at any given location within a scene, numerous visual fi lters are likely to be simultaneously active. Thus, the output of any given fi lter responding to real-world content will be weighted by the interdependent responses from a broad assortment of fi lters. ...
Article
Full-text available
Human contrast sensitivity for narrowband Gabor targets is suppressed when superimposed on narrowband masks of the same spatial frequency and orientation (referred to as overlay suppression), with suppression being broadly tuned to orientation and spatial frequency. Numerous behavioral and neurophysiological experiments have suggested that overlay suppression originates from the initial LGN inputs to V1, which is consistent with the broad tuning typically reported for overlay suppression. However, recent reports have shown narrowly tuned anisotropic overlay suppression when narrowband targets are masked by broadband noise. Consequently, researchers have argued for an additional form of overlay suppression that involves cortical contrast gain control processes. The current study sought to further explore this notion behaviorally using narrowband and broadband masks, along with a computational neural simulation of the hypothesized underlying gain control processes in cortex. Additionally, we employed transcranial direct current stimulation (tDCS) in order to test whether cortical processes are involved in driving narrowly tuned anisotropic suppression. The behavioral results yielded anisotropic overlay suppression for both broadband and narrowband masks and could be replicated with our computational neural simulation of anisotropic gain control. Further, the anisotropic form of overlay suppression could be directly modulated by tDCS, which would not be expected if the suppression was primarily subcortical in origin. Altogether, the results of the current study provide further evidence in support of an additional overlay suppression process that originates in cortex, and show that this form of suppression is also observable with narrowband masks.
Preprint
Full-text available
Decision confidence plays a critical role in humans' ability to make adaptive decisions in a noisy perceptual world. Despite its importance, there is currently little consensus about the computations underlying confidence judgements in perceptual decisions. One leading theory suggests that confidence is computed following the rules of Bayesian inference. Accordingly, the goal of the current study was to investigate a fundamental assumption of Bayesian models: the use of prior knowledge in subjective confidence. Rather than requiring participants to internalise the parameters of an arbitrary prior distribution, we capitalised on the existing probability distributions of features in natural scenes, which are known to play a critical role in guiding perception. Participants reported the subjective upright of naturalistic image target patches, and then reported their confidence in their orientation responses. We used computational modelling to relate the statistics of the targets to participants' responses, confirming that participants used the prior probability distribution of features in natural scenes to judge subjective upright. Critically, our results reveal that participants also used natural image priors to inform their confidence judgements. Our findings provide important evidence supporting a Bayesian characterisation of confidence and highlight the influence of environmental priors on confidence.
Article
Preview the book chapter: https://books.google.com.sg/books?id=TFT7DwAAQBAJ&pg=PA1&lpg=PA1&dq=yap+tiong+peng&source=bl&ots=3s4sAJq5nE&sig=ACfU3U3wmIM11Kz7XHzSXlFHzJ4nKVivKA&hl=en&sa=X&ved=2ahUKEwjzi77J18XzAhXVfX0KHVa_Ddk4RhDoAXoECAcQAg#v=onepage&q=yap%20tiong%20peng&f=false
Article
Orientation processing in visual cortex appears matched to the environment, such that larger neural populations are tuned to cardinal (horizontal/vertical) than oblique orientations. This may be manifested perceptually as a cardinal bias: poorer sensitivity to oblique compared to cardinal orientations (the “oblique effect”). However, a growing body of psychophysical data reveals the opposite pattern of anisotropy: a bias towards the oblique over the cardinal orientations (the “horizontal effect”), something matched by recent functional magnetic resonance imaging (fMRI) studies that have found an increased response to the oblique over the cardinal orientations in early visual cortex. This may reveal the operation of an efficient coding strategy optimised to the diet of orientations encountered during natural viewing. From consideration of coding efficiency, it might be expected that the anisotropieswould change as the quality/strength of the oriented stimulus changes. In two experiments, fMRI response modulations were measured in retinotopically-defined human early visual cortex as a function of the contrast and orientation of sinusoidal gratings. Both experiments revealed a marked change in the V1 response from a cardinal (vertical) bias at low contrast to an oblique bias at high contrast. In Experiment 2, this was also apparent in areas V2 and V3. On average, there was no systematic “radial bias” (a preference for orientations aligned with the visual field meridian) in V1, although it was present in some individual subjects. The change in orientation anisotropies with contrast is consistent with an adaptive stimulus coding strategy in cortex that shifts according to the strength of the sensory inputs.
Article
Full-text available
Real-world visual scenes display consistent first- and second-order statistical regularities to which visual neural representations may be perceptually matched, but these lower-order regularities stem from constraints on image power spectra, which appear to carry much less perceptual information than image phase spectra. Natural scenes are shown also to display consistent higher-order statistical regularities, and an analysis of these regularities in terms of fourth-order spectra shows that they are strongly dependent on spatial frequency. These findings have important consequences for the design of a visual system that aims to maximize sparseness in neural representations.
Article
Full-text available
In order to transmit information in images efficiently, the visual system should be tuned to the statistical structure of the ensemble of images that it sees. Several authors have suggested that the ensemble of natural images exhibits fractal behavior and, therefore, has a power spectrum that drops off proportionally to 1/fβ(2 < β < 4). In this paper we investigate the question of which value of the exponent β describes the power spectrum of the ensemble of images to which the visual system is optimally tuned. An experiment in which subjects were asked to discriminate randomly generated noise textures based on their spectral drop-off was used. Whereas the discrimination-threshold function of an ideal observer was flat for different spectral drop-offs, human observers showed a broad peak in sensitivity for 2.8 < β < 3.6. The results are consistent with, but do not provide direct evidence for, the theory that the visual system is tuned to an ensemble of images with Markov statistics.
Article
Full-text available
Inferior performance for obliquely oriented stimuli is often observed on higher level visual and somatosensory tasks and also on tests of low-level visual sensory ability. This study demonstrated an anisotropy of low-level somatosensory performance. Sensitivity to gratings on the finger pad was highest for gratings oriented proximally-distally, intermediate for oblique gratings, and lowest for medial-lateral gratings. This pattern supports a model proposing that detection threshold is determined by the number of neurons tuned to a stimulus (A. Anzai, A. Bearse, Jr., R. D. Freeman, & D. Cai, 1995). A classification of somatosensory and visual anisotropies is proposed in which orientation biases are classified as being attributable to either anisotropic sensory filtering (Class 1) or anisotropic higher level processing (Class 2). It was concluded that a given instance of anisotropic visual or somatosensory performance may stem from low-level sensory factors, high-level factors, or a mixture of the two, depending on the task demands. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Full-text available
Several studies of natural-image statistics have concluded that the majority of structural image information is represented in the Fourier domain not by the power spectrum, which can be modeled with a power function of spatial frequency, but by the phase spectrum. Psychophysical experiments are described in which human observers were required to discriminate between visual stimuli differing only in their global second-order statistics. The patterns of sensitivity recorded were found to depend on whether these stimuli were natural images or images with random-phase spectra. These results suggest that discrimination performance is related to changes in higher-order image structure; a bandpass model observer based on third-moment statistics was applied to all stimulus images and was found to provide a good fit to the psychophysical data.
Article
Physiologists have long proposed that correlated input activity is important in normal sensory development. Here it is postulated that the visual system is sensitive to the correlation in image intensity across the visual field, and that these correlations are used to help calibrate spatial representations. Since measurements made near to each other in the visual field are more correlated than measurements made at a distance, the degree of correlation can be used as an estimate of the distance between two measurements and can therefore be used to calibrate a roughly organized spatial representation. We therefore explored the hypothesis that low level spatial representations are calibrated using a signal based on image intensity correlation. If the visual system uses input statistics to calibrate its spatial representation, then any distortions and anisotropies in these input statistics should be mirrored by distortions in the representation of space. To test the psychological implications of this hypothesis, a collection of 81 images of open and urban landscapes were used to estimate the degree of correlation between image intensity measurement pairs as a function of both distance and orientation. Doing this we show that a system that used the statistics measured to calibrate its representation would show: (1) a horizontal-vertical illusion;(2) the magnitude of this illusion would depend on the amount of open and urban landscapes in the environment;(3) there would be a nontrivial relationship between line orientation and judged length. Analogues of all these distortions and regularities can be found in the psychophysical literature on distance estimation. This gives strength to the proposal that spatial representations are calibrated using input statistics.
Article
Measurements have been made of some basic statistical quantities characterizing picture signals. These include various amplitude distributions, autocorrelation, and correlation among successive frames. The methods of measurement are described, and the results are used to estimate the amount by which the channel capacity required for television transmission may be reduced through exploitation of the statistics measured.
Article
This study is concerned with an analysis of the properties of the stimulating situation which form the basis for discrimination of visual objects. The apparatus used required the animal to jump against stimulus cards from a distance of 20 cm. with the reward of food for a correct choice and punishment by a fall into a net for an incorrect one. Evidence was obtained that the figure-ground relation exists in the animal's visual field and that the segregation of figure is determined by at least some of the same factors which are effective in human vision. Discrimination was found to be based upon the abstraction of certain general properties of the figure, which are then recognized in non-identical figures. Ease of discrimination depends upon the presence of certain relational properties, such as predominant direction, in the figures. Differences in the ease of discrimination and recognition of different figures indicate that the rat's visual system functions most efficiently in spatial orientation and that the identification of objects or forms is secondary to a system of space coordinates. The bearing of the observations upon the general problem of the nature and evolution of intelligence is discussed. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Sensitivity for vertical and horizontal gratings is higher than for obliquely oriented gratings, a phenomenon known as the oblique effect. We compared visual evoken potential and preferential looking estimates of the oblique effect in nonastigmatic 3-month-old infants and found that orientation effects occurred more frequently by visual evoked potential measure than by preferential looking.
Article
Most natural images, although apparently diverse, have surprisingly regular 1/fβ Fourier amplitude spectra (average β: 1.05–1.20). If the visual system evolved and develops in this spectral environment, is there a sense in which human vision displays commensurate fractal properties? Correlation analysis was applied to human spatial frequency contrast sensitivity. Adults — but not infants — show power law correlational structure consistent with a β of 1.09–1.20 (closely matched to that of natural images). This result provides insights into multi-scale spatial frequency processing in human vision and provides a basis for understanding perceptual phenomena like contrast control and motion deblurring.
Article
Optical Fourier analysis of photographic samples of three visual environments—indoor carpentered, outdoor carpentered, and pastoral—showed that in the 1–25 c/deg spatial frequency range, carpentered environments contained more information in horizontal (H) and vertical (V) orientations. However, in the 5–25 c/deg range the V orientation dominated, and the pastoral environment had the greatest anisotropy. Thus, a spatial-frequency-specific influence of carpentered environments on anisotropic acuity, which favors both H and V at higher frequencies, is not substantiated.