Facilitative and inhibitory factors in visual texture discrimination

In this paper we consider some spatial and temporal properties of visual textures which either inhibit or enhance their discrimination. From exposure time and texture mixing studies two findings emerge. First, it is clear that the spatial summation process involved in texture discrimination has a time course consistent with counting models for decision making. Secondly, the receptive field structures of salient texture features, in particular-orientation, seem to involve simple excitatory center and inhibitory surround mechanisms.This project was supported by a grant from the Australian Research Grant Council. I would like to thank Ms. Judy McKimm for aiding in the collection of experimental data     

Asymmetries in visual texture discrimination

Recent results have shown that texture discrimination is an asymmetrical process; texture A within texture B may be much easier to detect than texture B within texture A. Two questions regarding discrimination asymmetries are addressed: (i) what sorts of textural properties are associated with discrimination asymmetries; and (ii) what sort of architecture would yield asymmetries. Two experiments show that discrimination asymmetries obtain when textures comprise circles of different sizes (large circles are easier to detect in small than vice versa) and when circles differ only in the regularity of their placement (irregularly placed circles are easier to detect in a background of regularly placed circles than vice versa). A plausible account of texture discrimination would involve the decomposition of images via a set orientation and scale selective filters followed by a second layer of filtering to detect energy differences between adjacent regions in the original convolutions. Discrimination asymmetries provide prima facie evidence against such a model because it involves only local measurements and comparisons. We propose that discrimination asymmetries are elegantly explained if it is assumed that the responses of the orientation and scale selective filters are normalized by the degree to which similarly tuned operators are responding elsewhere in the image; viz., global normalization of filter responses. However, there are cases where such global normalization is not required to explain asymmetrical discrimination.     

Receptive fields in the rat piriform cortex.

Current models of odor discrimination in mammals involve molecular feature detection by a large family of diverse olfactory receptors, refinement of molecular feature extraction through precise projections of olfactory receptor neurons to the olfactory bulb to form an odor-specific spatial map of molecular features across glomerular layer, and synthesis of these features into odor objects within the piriform cortex. This review describes our recent work on odor and spatial receptive fields within the anterior piriform cortex and compares these fields with receptive fields of their primary afferent, olfactory bulb mitral/tufted cells. The results suggest that receptive fields in the piriform cortex are ensemble in nature, highly dynamic, and may contribute to odor discrimination and odor memory.     

Discrimination of complex textures by bees

A problem confronted by visual systems is that of discriminating textures. It appears that a recently described class of orientation-tuned neurones in the bee brain embody properties of mechanisms used by humans to discriminate complex textures. In particular these mechanisms would permit bees to discriminate a large range of textures by giving bees access to information related to higher-order correlations between texture elements. To determine if bees can exploit such textural information we have conducted behavioural experiments employing iso-dipole textures, that statistically speaking, differ from binary noise textures, and each other, only in their third-order correlation functions. While these textures are not themselves of any ethological significance their special properties permit us to show that bees can potentially use a very large palette of textures to classify textured objects. In electrophysiological experiments we demonstrate the requisite contrast sign invariance (rectification) of the orientation-selective neurones' responses and discuss other similarities of these neurones' responses to models accounting for human texture discrimination. Accepted: 7 October 1998     

Idealized Computational Models for Auditory Receptive Fields

We present a theory by which idealized models of auditory receptive fields can be derived in a principled axiomatic manner, from a set of structural properties to (i) enable invariance of receptive field responses under natural sound transformations and (ii) ensure internal consistency between spectro-temporal receptive fields at different temporal and spectral scales. For defining a time-frequency transformation of a purely temporal sound signal, it is shown that the framework allows for a new way of deriving the Gabor and Gammatone filters as well as a novel family of generalized Gammatone filters, with additional degrees of freedom to obtain different trade-offs between the spectral selectivity and the temporal delay of time-causal temporal window functions. When applied to the definition of a second-layer of receptive fields from a spectrogram, it is shown that the framework leads to two canonical families of spectro-temporal receptive fields, in terms of spectro-temporal derivatives of either spectro-temporal Gaussian kernels for non-causal time or a cascade of time-causal first-order integrators over the temporal domain and a Gaussian filter over the logspectral domain. For each filter family, the spectro-temporal receptive fields can be either separable over the time-frequency domain or be adapted to local glissando transformations that represent variations in logarithmic frequencies over time. Within each domain of either non-causal or time-causal time, these receptive field families are derived by uniqueness from the assumptions. It is demonstrated how the presented framework allows for computation of basic auditory features for audio processing and that it leads to predictions about auditory receptive fields with good qualitative similarity to biological receptive fields measured in the inferior colliculus (ICC) and primary auditory cortex (A1) of mammals.     

Tuned solutions in dynamic neural fields as building blocks for extended EEG models

The most prominent functional property of cortical neurons in sensory areas are their tuned receptive fields which provide specific responses of the neurons to external stimuli. Tuned neural firing indeed reflects the most basic and best worked out level of cognitive representations. Tuning properties can be dynamic on a short time-scale of fractions of a second. Such dynamic effects have been modeled by localised solutions (also called “bumps” or “peaks”) in dynamic neural fields. In the present work we develop an approximation method to reduce the dynamics of localised activation peaks in systems of n coupled nonlinear d-dimensional neural fields with transmission delays to a small set of delay differential equations for the peak amplitudes and widths only. The method considerably simplifies the analysis of peaked solutions as demonstrated for a two-dimensional example model of neural feature selectivity in the brain. The reduced equations describe the effective interaction between pools of local neurons of several (n) classes that participate in shaping the dynamic receptive field responses. To lowest order they resemble neural mass models as they often form the base of EEG-models. Thereby they provide a link between functional small-scale receptive field models and more coarse-grained EEG-models. More specifically, they connect the dynamics in feature-selective cortical microcircuits to the more abstract local elements used in coarse-grained models. However, beside amplitudes the reduced equations also reflect the sharpness of tuning of the activity in a d-dimensional feature space in response to localised stimuli.     

The Visual N1 Is Sensitive to Deviations from Natural Texture Appearance

Disruptions of natural texture appearance are known to negatively impact performance in texture discrimination tasks, for example, such that contrast-negated textures, synthetic textures, and textures depicting abstract art are processed less efficiently than natural textures. Presently, we examined how visual ERP responses (the P1 and the N1 in particular) were affected by violations of natural texture appearance. We presented participants with images depicting either natural textures or synthetic textures made from the original stimuli. Both stimulus types were additionally rendered either in positive or negative contrast. These appearance manipulations (negation and texture synthesis) preserve a range of low-level features, but also disrupt higher-order aspects of texture appearance. We recorded continuous EEG while participants completed a same/different image discrimination task using these images and measured both the P1 and N1 components over occipital recording sites. While the P1 exhibited no sensitivity to either contrast polarity or real/synthetic appearance, the N1 was sensitive to both deviations from natural appearance. Polarity reversal and synthetic appearance affected the N1 latency differently, however, suggesting a differential impact on processing. Our results suggest that stages of visual processing indexed by the P1 and N1 are sensitive to high-order statistical regularities in natural textures and also suggest that distinct violations of natural appearance impact neural responses differently.     

Local spectral analysis in the visual cortex

On the basis of experimental evidence presented earlier a model of local spectral analysis of the image performed by the complex receptive fields of the visual cortex has been proposed. An essential feature of the model is that the generalized piece-wise Fourier transform is performed not over the image luminance function but over the logarithm contrast function resulted from analysis of the image by the round receptive fields of the preceding levels. Such an assumption removes a number of experimental objections offered against the hypothesis of two-dimensional Fourier transform in the visual system. The consequencies from the piece-wise expansion in a series of basic functions have been considered and among them: the channel frequency characteristics which can have more than one maximum; the possibility of describing the image by a limited number of channels with overlapping frequency characteristics; the existence of mechanisms for estimation of phase shift between frequencies.     

Illuminant precompensation for texture discrimination using filters

 Texture-discrimination algorithms have often been tested on images containing either mosaics of synthetic textures or artificially created mosaics of real textures – in any case, images in which most of the changes in intensity can be ascribed to the textures themselves. However, real images are not formed like this and may contain steep gradations in intensity which have nothing to do with local texture, such as those caused by incident shadows. A texture discrimination algorithm based on linear filters can fail in the presence of these strong gradations, as they may easily contain an order of magnitude more energy than the gradations in intensity due to texture in the image per se. In these cases, the mechanism may become responsive only to strong luminance effects, and not to texture. I have found that good performance on natural images containing texture can only be obtained from a filter-based texture detection scheme if it includes a stage which attempts to bring large intensity gradients within bounds. The exact nature of the best precompensator appears to depend somewhat on the way the filter outputs are processed. The fit to psychophysical data and the implications for more detailed models of human texture processing will be discussed. Received: 3 May 1993/Accepted in revised form: 7 June 1993     

Synergy of features enables detection of texture defined figures

Traditional theories of early visual processing suggest that elementary visual features are handled in parallel by independent neural pathways. We studied the interaction of orientation and spatial frequency in the discrimination of Gabor random fields. Target textures differed from reference textures either in mean feature value, showing an edge-like transition between both textures (edge defined), or in the degree of feature homogeneity with smooth transitions (region defined). Irrespective of the kind of texture definition, we found strong cue summation for targets defined by both cues simultaneously, provided two conditions were fulfilled. First, they were barely discriminable when defined by one cue alone. Second, the target elements formed a closed 2D surface. Only marginal cue summation was observed when target elements were heterogeneously distributed in a predefined area, lacking a clear 2D shape. Our findings indicate that feature synergy enables figure-ground segregation when the information from independent feature-specific pathways is insufficient for solving this task.     

On perceptual analyzers underlying visual texture discrimination: Part II

In Part I Caelli and Julesz generated texture pairs of 4-disk micropatterns with identical dipole statistics. They found that this iso-dipole constraint could not prevent the quasi-collinearity of certain disk elements which, in turn, yielded effortless discrimination. They proposed two classes of perceptual analyzers to explain discrimination with these micropatern textures: Class A, corresponding to those which detect dipole differences; while Class B detectors, such as the quasi-collinear detector (QCD), acted when isodipole textures were presented. In this paper we show several new methods for generating iso-dipole textures with micropatterns consisting of 5 or more disks or non-disk shaped elements, and we report the discovery of two other Class B detectors, a corner detector (using a 6-disk method), and a closure detector (with 8–11 disk micropatterns). These QCD, corner, and closure detectors were verified by examining several hundred iso-dipole texture pairs. It appears that iso-dipole constraints make ineffective all other feature analyzers involved in effortless texture discrimination than the Class B types. These figural properties of collinearity, corners, and closure can be perceived without scrutiny and are precursors of form perception.     

Texture discrimination by Gabor functions

A 2D Gabor filter can be realized as a sinusoidal plane wave of some frequency and orientation within a two dimensional Gaussian envelope. Its spatial extent, frequency and orientation preferences as well as bandwidths are easily controlled by the parameters used in generating the filters. However, there is an uncertainty relation associated with linear filters which limits the resolution simultaneously attainable in space and frequency. Daugman (1985) has determined that 2D Gabor filters are members of a class of functions achieving optimal joint resolution in the 2D space and 2D frequency domains. They have also been found to be a good model for two dimensional receptive fields of simple cells in the striate cortex (Jones 1985; Jones et al. 1985).The characteristic of optimal joint resolution in both space and frequency suggests that these filters are appropriate operators for tasks requiring simultaneous measurement in these domains. Texture discrimination is such a task. Computer application of a set of Gabor filters to a variety of textures found to be preattentively discriminable produces results in which differently textured regions are distinguished by firstorder differences in the values measured by the filters. This ability to reduce the statistical complexity distinguishing differently textured region as well as the sensitivity of these filters to certain types of local features suggest that Gabor functions can act as detectors of certain texton types. The performance of the computer models suggests that cortical neurons with Gabor like receptive fields may be involved in preattentive texture discrimination.     

Simple model for encoding natural images by retinal ganglion cells with nonlinear spatial integration

A central goal in sensory neuroscience is to understand the neuronal signal processing involved in the encoding of natural stimuli. A critical step towards this goal is the development of successful computational encoding models. For ganglion cells in the vertebrate retina, the development of satisfactory models for responses to natural visual scenes is an ongoing challenge. Standard models typically apply linear integration of visual stimuli over space, yet many ganglion cells are known to show nonlinear spatial integration, in particular when stimulated with contrast-reversing gratings. We here study the influence of spatial nonlinearities in the encoding of natural images by ganglion cells, using multielectrode-array recordings from isolated salamander and mouse retinas. We assess how responses to natural images depend on first- and second-order statistics of spatial patterns inside the receptive field. This leads us to a simple extension of current standard ganglion cell models. We show that taking not only the weighted average of light intensity inside the receptive field into account but also its variance over space can partly account for nonlinear integration and substantially improve response predictions of responses to novel images. For salamander ganglion cells, we find that response predictions for cell classes with large receptive fields profit most from including spatial contrast information. Finally, we demonstrate how this model framework can be used to assess the spatial scale of nonlinear integration. Our results underscore that nonlinear spatial stimulus integration translates to stimulation with natural images. Furthermore, the introduced model framework provides a simple, yet powerful extension of standard models and may serve as a benchmark for the development of more detailed models of the nonlinear structure of receptive fields.     

Are V1 Simple Cells Optimized for Visual Occlusions? A Comparative Study

Simple cells in primary visual cortex were famously found to respond to low-level image components such as edges. Sparse coding and independent component analysis (ICA) emerged as the standard computational models for simple cell coding because they linked their receptive fields to the statistics of visual stimuli. However, a salient feature of image statistics, occlusions of image components, is not considered by these models. Here we ask if occlusions have an effect on the predicted shapes of simple cell receptive fields. We use a comparative approach to answer this question and investigate two models for simple cells: a standard linear model and an occlusive model. For both models we simultaneously estimate optimal receptive fields, sparsity and stimulus noise. The two models are identical except for their component superposition assumption. We find the image encoding and receptive fields predicted by the models to differ significantly. While both models predict many Gabor-like fields, the occlusive model predicts a much sparser encoding and high percentages of ‘globular’ receptive fields. This relatively new center-surround type of simple cell response is observed since reverse correlation is used in experimental studies. While high percentages of ‘globular’ fields can be obtained using specific choices of sparsity and overcompleteness in linear sparse coding, no or only low proportions are reported in the vast majority of studies on linear models (including all ICA models). Likewise, for the here investigated linear model and optimal sparsity, only low proportions of ‘globular’ fields are observed. In comparison, the occlusive model robustly infers high proportions and can match the experimentally observed high proportions of ‘globular’ fields well. Our computational study, therefore, suggests that ‘globular’ fields may be evidence for an optimal encoding of visual occlusions in primary visual cortex.     

Strength of feature contrast mediates interaction among feature domains

Traditional theories of texture segregation suggest that elementary visual features are processed in parallel by independent modules at early visual stages. Here we show that, for small feature contrasts and large values evoking perceptual popout, different forms of module interaction exist. While discrimination of highly salient features rests on independent feature specific pathways, information is summed across domains when barely noticeable ones are to be detected in homogeneous textures.     

Pattern-dependent response modulations in motion-sensitive visual interneurons--a model study

Even if a stimulus pattern moves at a constant velocity across the receptive field of motion-sensitive neurons, such as lobula plate tangential cells (LPTCs) of flies, the response amplitude modulates over time. The amplitude of these response modulations is related to local pattern properties of the moving retinal image. On the one hand, pattern-dependent response modulations have previously been interpreted as 'pattern-noise', because they deteriorate the neuron's ability to provide unambiguous velocity information. On the other hand, these modulations might also provide the system with valuable information about the textural properties of the environment. We analyzed the influence of the size and shape of receptive fields by simulations of four versions of LPTC models consisting of arrays of elementary motion detectors of the correlation type (EMDs). These models have previously been suggested to account for many aspects of LPTC response properties. Pattern-dependent response modulations decrease with an increasing number of EMDs included in the receptive field of the LPTC models, since spatial changes within the visual field are smoothed out by the summation of spatially displaced EMD responses. This effect depends on the shape of the receptive field, being the more pronounced--for a given total size--the more elongated the receptive field is along the direction of motion. Large elongated receptive fields improve the quality of velocity signals. However, if motion signals need to be localized the velocity coding is only poor but the signal provides--potentially useful--local pattern information. These modelling results suggest that motion vision by correlation type movement detectors is subject to uncertainty: you cannot obtain both an unambiguous and a localized velocity signal from the output of a single cell. Hence, the size and shape of receptive fields of motion sensitive neurons should be matched to their potential computational task.     

Invariance and selectivity in the ventral visual pathway.

Pattern recognition systems that are invariant to shape, pose, lighting and texture are never sufficiently selective; they suffer a high rate of "false alarms". How are biological vision systems both invariant and selective? Specifically, how are proper arrangements of sub-patterns distinguished from the chance arrangements that defeat selectivity in artificial systems? The answer may lie in the nonlinear dynamics that characterize complex and other invariant cell types: these cells are temporarily more receptive to some inputs than to others (functional connectivity). One consequence is that pairs of such cells with overlapping receptive fields will possess a related property that might be termed functional common input. Functional common input would induce high correlation exactly when there is a match in the sub-patterns appearing in the overlapping receptive fields. These correlations, possibly expressed as a partial and highly local synchrony, would preserve the selectivity otherwise lost to invariance.