Computational model for perception of objects and motions

Perception of objects and motions in the visual scene is one of the basic problems in the visual system. There exist 'What' and 'Where' pathways in the superior visual cortex, starting from the simple cells in the primary visual cortex. The former is able to perceive objects such as forms, color, and texture, and the latter perceives 'where', for example, velocity and direction of spatial movement of objects. This paper explores brain-like computational architectures of visual information processing. We propose a visual perceptual model and computational mechanism for training the perceptual model. The compu- tational model is a three-layer network. The first layer is the input layer which is used to receive the stimuli from natural environments. The second layer is designed for representing the internal neural information. The connections between the first layer and the second layer, called the receptive fields of neurons, are self-adaptively learned based on principle of sparse neural representation. To this end, we introduce Kullback-Leibler divergence as the measure of independence between neural responses and derive the learning algorithm based on minimizing the cost function. The proposed algorithm is applied to train the basis functions, namely receptive fields, which are localized, oriented, and bandpassed. The resultant receptive fields of neurons in the second layer have the characteristics resembling that of simple cells in the primary visual cortex. Based on these basis functions, we further construct the third layer for perception of what and where in the superior visual cortex. The proposed model is able to perceive objects and their motions with a high accuracy and strong robustness against additive noise. Computer simulation results in the final section show the feasibility of the proposed perceptual model and high efficiency of the learning algorithm.     

Characteristics of Human Luminance Discrimination and Modeling a Neural Network Based on the Response Properties of the Visual Cortex

Reaction time (RT) and error rate that depend on stimulus duration were measured in a luminance-discrimination reaction time task. Two patches of light with different luminance were presented to participants for ‘short’ (150 ms) or ‘long’ (1 s) period on each trial. When the stimulus duration was ‘short’, the participants responded more rapidly with poorer discrimination performance than they did in the longer duration. The results suggested that different sensory responses in the visual cortices were responsible for the dependence of response speed and accuracy on the stimulus duration during the luminance-discrimination reaction time task. It was shown that the simple winner-take-all-type neural network model receiving transient and sustained stimulus information from the primary visual cortex successfully reproduced RT distributions for correct responses and error rates. Moreover, temporal spike sequences obtained from the model network closely resembled to the neural activity in the monkey prefrontal or parietal area during other visual decision tasks such as motion discrimination and oddball detection tasks.     

Fast cortical selection: a principle of neuronal self-organization for perception?

 It is commonly accepted that larger visual objects are represented in the cerebral cortex by specific spatial patterns of neuronal activity. Self-organization is a key concept in the different explanations of such neuronal representations. We here propose as a hypothesis that fast cortical selection (FCS) is an intrinsic functional element of cortical self-organization during perception. Selection is a central concept in theoretical biology which has proved its explanatory power in different fields of our natural and cultural world. The central element in the cortical selection process is the pyramidal cell with its two types of excitatory input. In primary cortical areas one of these inputs comes from any of the sensory organs, determining the topological and typological receptive field properties of the cell and also driving it directly. The other type of input connects reciprocally neighbouring pyramidal cells by axon collaterals and only facilitates the driving input. These two functionally different inputs constitute the elementary selection system working by iterative mutual facilitation as a biological algorithm. A short simulation, based entirely on such biological facts, illustrates the dynamic of this selection process: the activity of cells responding better to the external stimulus ‘grow and survive’ the stimulation, whereas less responsive cells decrease their activity due to competition. Received: 13 June 1995 / Accepted in revised form: 27 May 1997     

The Gaussian derivative model for spatial-temporal vision: I. Cortical model

How do we see the motion of objects as well as their shapes? The Gaussian Derivative (GD) spatial model is extended to time to help answer this question. The GD spatio-temporal model requires only two numbers to describe the complete three-dimensional space-time shapes of individual receptive fields in primate visual cortex. These two numbers are the derivative numbers along the respective spatial and temporal principal axes of a given receptive field. Nine transformation parameters allow for a standard geometric association of these intrinsic axes with the extrinsic environment. The GD spatio-temporal model describes in one framework the following properties of primate simple cell fields: motion properties, number of lobes in space-time, spatial orientation. location, and size. A discrete difference-of-offset-Gaussians (DOOG) model provides a plausible physiological mechanism to form GD-like model fields in both space and time. The GD model hypothesizes that receptive fields at the first stage of processing in the visual cortex approximate 'derivative analyzers' that estimate local spatial and temporal derivatives of the intensity profile in the visual environment. The receptive fields as modeled provide operators that can allow later stages of processing in either a biological or machine vision system to estimate the motion as well as the shapes of objects in the environment.     

Neural network model of visual cortex for determining surface curvature from images of shaded surfaces

The visual system can extract information about shape from the pattern of light and dark surface shading on an object. Very little is known about how this is accomplished. We have used a learning algorithm to construct a neural network model that computes the principal curvatures and orientation of elliptic paraboloids independently of the illumination direction. Our chief finding is that receptive fields developed by units of such model network are surprisingly similar to some found in the visual cortex. It appears that neurons that can make use of the continuous gradations of shading have receptive fields similar to those previously interpreted as dealing with contours (i.e. 'bar' detectors or 'edge' detectors). This study illustrates the difficulty of deducing neuronal function within a network solely from receptive fields. It is also important to consider the pattern of connections a neuron makes with subsequent stages, which we call the 'projective field'.     

Learning receptive fields using predictive feedback.

Previously, it was suggested that feedback connections from higher- to lower-level areas carry predictions of lower-level neural activities, whereas feedforward connections carry the residual error between the predictions and the actual lower-level activities [Rao, R.P.N., Ballard, D.H., 1999. Nature Neuroscience 2, 79-87.]. A computational model implementing the hypothesis learned simple cell receptive fields when exposed to natural images. Here, we use predictive feedback to explain tuning properties in medial superior temporal area (MST). We implement the hypothesis using a new, biologically plausible, algorithm based on matching pursuit, which retains all the features of the previous implementation, including its ability to efficiently encode input. When presented with natural images, the model developed receptive field properties as found in primary visual cortex. In addition, when exposed to visual motion input resulting from movements through space, the model learned receptive field properties resembling those in MST. These results corroborate the idea that predictive feedback is a general principle used by the visual system to efficiently encode natural input.     

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.     

Perisaccadic Remapping and Rescaling of Visual Responses in Macaque Superior Colliculus

Visual neurons have spatial receptive fields that encode the positions of objects relative to the fovea. Because foveate animals execute frequent saccadic eye movements, this position information is constantly changing, even though the visual world is generally stationary. Interestingly, visual receptive fields in many brain regions have been found to exhibit changes in strength, size, or position around the time of each saccade, and these changes have often been suggested to be involved in the maintenance of perceptual stability. Crucial to the circuitry underlying perisaccadic changes in visual receptive fields is the superior colliculus (SC), a brainstem structure responsible for integrating visual and oculomotor signals. In this work we have studied the time-course of receptive field changes in the SC. We find that the distribution of the latencies of SC responses to stimuli placed outside the fixation receptive field is bimodal: The first mode is comprised of early responses that are temporally locked to the onset of the visual probe stimulus and stronger for probes placed closer to the classical receptive field. We suggest that such responses are therefore consistent with a perisaccadic rescaling, or enhancement, of weak visual responses within a fixed spatial receptive field. The second mode is more similar to the remapping that has been reported in the cortex, as responses are time-locked to saccade onset and stronger for stimuli placed in the postsaccadic receptive field location. We suggest that these two temporal phases of spatial updating may represent different sources of input to the SC.     

Objects limit human comprehension

This paper demonstrates that the human visual system, the primary sensory conduit for primates, processes ambient energy in a way that obligatorily constructs the objects that we ineluctably perceive. And since our perceptual apparatus processes information only in terms of objects (along with the properties and movements of objects), we are limited in our ability to comprehend ‘what is’ when we move beyond our ordinary world of midsize objects—as, for example, when we address the micro microworld of quantum physics.

Prior expectations evoke stimulus-specific activity in the deep layers of the primary visual cortex

The way we perceive the world is strongly influenced by our expectations. In line with this, much recent research has revealed that prior expectations strongly modulate sensory processing. However, the neural circuitry through which the brain integrates external sensory inputs with internal expectation signals remains unknown. In order to understand the computational architecture of the cortex, we need to investigate the way these signals flow through the cortical layers. This is crucial because the different cortical layers have distinct intra- and interregional connectivity patterns, and therefore determining which layers are involved in a cortical computation can inform us on the sources and targets of these signals. Here, we used ultra-high field (7T) functional magnetic resonance imaging (fMRI) to reveal that prior expectations evoke stimulus-specific activity selectively in the deep layers of the primary visual cortex (V1). These findings are in line with predictive processing theories proposing that neurons in the deep cortical layers represent perceptual hypotheses and thereby shed light on the computational architecture of cortex.

The way we perceive the world is strongly influenced by our expectations, but the neural circuitry through which the brain achieves this remains unknown. A study using ultra-high field fMRI reveals that prior expectations evoke stimulus-specific signals in the deep layers of the primary visual cortex.     

Neural processes in symmetry perception: a parallel spatio-temporal model

Symmetry is usually computationally expensive to detect reliably, while it is relatively easy to perceive. In spite of many attempts to understand the neurofunctional properties of symmetry processing, no symmetry-specific activation was found in earlier cortical areas. Psychophysical evidence relating to the processing mechanisms suggests that the basic processes of symmetry perception would not perform a serial, point-by-point comparison of structural features but rather operate in parallel. Here, modeling of neural processes in psychophysical detection of bilateral texture symmetry is considered. A simple fine-grained algorithm that is capable of performing symmetry estimation without explicit comparison of remote elements is introduced. A computational model of symmetry perception is then described to characterize the underlying mechanisms as one-dimensional spatio-temporal neural processes, each of which is mediated by intracellular horizontal connections in primary visual cortex and adopts the proposed algorithm for the neural computation. Simulated experiments have been performed to show the efficiency and the dynamics of the model. Model and human performances are comparable for symmetry perception of intensity images. Interestingly, the responses of V1 neurons to propagation activities reflecting higher-order perceptual computations have been reported in neurophysiologic experiments.     

Adaptation of oriented and unoriented color-selective neurons in human visual areas

Primary visual cortex contains at least two distinct populations of color-selective cells: neurons in one have circularly symmetric receptive fields and respond best to reddish and greenish light, while neurons in another have oriented receptive fields and a variety of color preferences. The relative prevalence and perceptual roles of the two kinds of neurons remain controversial, however. We used fMRI and a selective adaptation technique to measure responses attributable to these two populations. The technique revealed evidence of adaptation in both populations and indicated that they each produced strong signals in V1 and other human visual areas. The activity of both sets of neurons was also reflected in color appearance measurements made with the same stimuli. Thus, both oriented and unoriented color-selective cells in V1 are important components of the neural pathways that underlie perception of color.     

A ring model for spatiotemporal properties of simple cells in the visual cortex

 A neural model is proposed for the spatiotemporal properties of simple cells in the visual cortex. In the model, several cortical cells are arranged on a ring, with mutual excitatory or inhibitory connections. The cells also receive excitatory inputs either from lagged and nonlagged cells of the lateral geniculate nucleus in one setting or from nonlagged cells in the other. Computer simulation shows that the cortical cells have spatiotemporally inseparable receptive fields in the former setting and separable fields in the latter; spatial profiles at a given time in the spatiotemporal fields are described with a Gabor function whose phase parameter varies regularly from 0 to 2π with rotation along the ring; the inseparable cells have directional selectivity as observed physiologically. Received: 13 November 1995 / Accepted in revised form: 1 July 1997     

Recognition of partly occluded patterns: a neural network model

 Human beings are often able to read a letter or word partly occluded by contaminating ink stains. However, if the stains are completely erased and the occluded areas of the letter are changed to white, we usually have difficulty in reading the letter. In this article I propose a hypothesis explaining why a pattern is easier to recognize when it is occluded by visible objects than by invisible opaque objects. A neural network model is constructed based on this hypothesis. The visual system extracts various visual features from the input pattern and then attempts to recognize it. If the occluding objects are not visible, the visual system will have difficulty in distinguishing which features are relevant to the original pattern and which are newly generated by the occlusion. If the occluding objects are visible, however, the visual system can easily discriminate between relevant and irrelevant features and recognize the occluded pattern correctly. The proposed model is an extended version of the neocognitron model. The activity of the feature-extracting cells whose receptive fields cover the occluding objects is suppressed in an early stage of the hierarchical network. Since the irrelevant features generated by the occlusion are thus eliminated, the model can recognize occluded patterns correctly, provided the occlusion is not so large as to prevent recognition even by human beings. Received: 21 February 2000 / Accepted in revised form: 11 September 2000     

Direct extraction of curvature-based metric shape from stereo by view-modulated receptive fields

Any computation of metric surface structure from horizontal disparities depends on the viewing geometry, and analysing this dependence allows us to narrow down the choice of viable schemes. For example, all depth-based or slant-based schemes (i.e. nearly all existing models) are found to be unrealistically sensitive to natural errors in vergence. Curvature-based schemes avoid these problems and require only moderate, more robust view-dependent corrections to yield local object shape, without any depth coding. This fits the fact that humans are strikingly insensitive to global depth but accurate in discriminating surface curvature. The latter also excludes coding only affine structure. In view of new adaptation results, our goal becomes to directly extract retinotopic fields of metric surface curvatures (i.e. avoiding intermediate disparity curvature).To find a robust neural realisation, we combine new exact analysis with basic neural and psychophysical constraints. Systematic, step-by-step ‘design’ leads to neural operators which employ a novel family of ‘dynamic’ receptive fields (RFs), tuned to specific (bi-)local disparity structure. The required RF family is dictated by the non-Euclidean geometry that we identify as inherent in cyclopean vision. The dynamic RF-subfield patterns are controlled via gain modulation by binocular vergence and version, and parameterised by a cell-specific tuning to slant. Our full characterisation of the neural operators invites a range of new neurophysiological tests. Regarding shape perception, the model inverts widely accepted interpretations: It predicts the various types of errors that have often been mistaken for evidence against metric shape extraction.     

Efficiency of neural networks for prediction of in vitro culture conditions and inoculum properties for optimum productivity

This study represents an ANN based computational scheming of physical, chemical and biological parameters at flask level for mass multiplication of plants through micropropagation using bioreactors of larger volumes. The optimal culture environment at small scale for Glycyrrhiza plant was predicted by using neural network approach in terms of pH and volume of growth medium per culture flask, incubation room temperature and month of inoculation along with inoculum properties in terms of inoculum size, fresh weight and number of explant per flask. This kind of study could be a model system in commercial propagation of various economically important plants in bioreactors using tissue culture technique. In present course of study the ANN was trained by implementing MATLAB neural network. A feed-forward back propagation type network was created for input vector (seven input elements), with single hidden layer (seven nodes) and one output unit in output layer. The ‘tansig’ and ‘purelin’ transfer functions were adapted for hidden and output layers respectively. The four training functions viz. traingda, trainrp, traincgf, traincgb were randomly selected to train four networks which further examined with available dataset. The efficiency of neural networks was concluded by the comparison of results obtained from this study with that of empirical data obtained from the detailed tissue culture experiments and designated as Target set (mean fresh weight biomass per culture flask after 40 days of in vitro culture duration). Efficiency of networks for better training initialization was judged on the basis of comparative analysis of ‘Mean Square Error at zero epoch’ for each network trained in which the least error at initial point was observed with trainrp followed by traincgb and traincgf. A comparative assessment between experimental target data range obtained from wet lab practice and all trained network output range for the efficiency of trained networks for least deviation from target range revealed the output range of network ‘trainrp’ was closest to the empirical target range while least comparison was worked out from network ‘traincgb’ which had output range more than the target decided and ultimately showed meaningless result.     

Does neuronal synchrony underlie visual feature grouping?

Previous research suggests that synchronous neural activity underlies perceptual grouping of visual image features. The generality of this mechanism is unclear, however, as previous studies have focused on pairs of neurons with overlapping or collinear receptive fields. By sampling more broadly and employing stimuli that contain partially occluded objects, we have conducted a more incisive test of the binding by synchrony hypothesis in area MT. We find that synchrony in spiking activity shows little dependence on feature grouping, whereas gamma band synchrony in field potentials can be significantly stronger when features are grouped. However, these changes in gamma band synchrony are small relative to the variability of synchrony across recording sites and do not provide a robust population signal for feature grouping. Moreover, these effects are reduced when stimulus differences nearby the receptive fields are eliminated using partial occlusion. Our findings suggest that synchrony does not constitute a general mechanism of visual feature binding.     

Binding and segmentation of multiple objects through neural oscillators inhibited by contour information

 Temporal correlation of neuronal activity has been suggested as a criterion for multiple object recognition. In this work, a two-dimensional network of simplified Wilson-Cowan oscillators is used to manage the binding and segmentation problem of a visual scene according to the connectedness Gestalt criterion. Binding is achieved via original coupling terms that link excitatory units to both excitatory and inhibitory units of adjacent neurons. These local coupling terms are time independent, i.e., they do not require Hebbian learning during the simulations. Segmentation is realized by a two-layer processing of the visual image. The first layer extracts all object contours from the image by means of “retinal cells” with an “on-center” receptive field. Information on contour is used to selectively inhibit Wilson-Cowan oscillators in the second layer, thus realizing a strong separation among neurons in different objects. Accidental synchronism between oscillations in different objects is prevented with the use of a global inhibitor, i.e., a global neuron that computes the overall activity in the Wilson-Cowan network and sends back an inhibitory signal. Simulations performed in a 50×50 neural grid with 21 different visual scenes (containing up to eight objects + background) with random initial conditions demonstrate that the network can correctly segment objects in almost 100% of cases using a single set of parameters, i.e., without the need to adjust parameters from one visual scene to the next. The network is robust with reference to dynamical noise superimposed on oscillatory neurons. Moreover, the network can segment both black objects on white background and vice versa and is able to deal with the problem of “fragmentation.” The main limitation of the network is its sensitivity to static noise superimposed on the objects. Overcoming this problem requires implementation of more robust mechanisms for contour enhancement in the first layer in agreement with mechanisms actually realized in the visual cortex. Received: 25 October 2001 / Accepted: 26 February 2003 / Published online: 20 May 2003 Correspondence to: Mauro Ursino (e-mail: mursino@deis.unibo.it, Tel.: +39-051-2093008, Fax: +39-051-2093073)     

Deriving behavioural receptive fields for visually completed contours

The visual system is constantly faced with the problem of identifying partially occluded objects from incomplete images cast on the retinae. Phenomenologically, the visual system seems to fill in missing information by interpolating illusory and occluded contours at points of occlusion, so that we perceive complete objects. Previous behavioural [1] [2] [3] [4] [5] [6] [7] and physiological [8] [9] [10] [11] [12] studies suggest that the visual system treats illusory and occluded contours like luminance-defined contours in many respects. None of these studies has, however, directly shown that illusory and occluded contours are actually used to perform perceptual tasks. Here, we use a response-classification technique [13] [14] [15] [16] [17] [18] [19] [20] to answer this question directly. This technique provides pictorial representations - 'classification images' - that show which parts of a stimulus observers use to make perceptual decisions, effectively deriving behavioural receptive fields. Here we show that illusory and occluded contours appear in observers' classification images, providing the first direct evidence that observers use perceptually interpolated contours to recognize objects. These results offer a compelling demonstration of how visual processing acts on completed representations, and illustrate a powerful new technique for constraining models of visual completion.     

Object segmentation model: analytical results and biological implications

 A simple, biologically motivated neural network for segmentation of a moving object from a visual scene is presented. The model consists of two parts: an object selection model which employs a scaling approach for receptive field sizes, and a subsequent network implementing a spotlight by means of multiplicative synapses. The network selects one object out of several, segments the rough contour of the object, and encodes the winner object's position with high accuracy. Analytical equations for the performance level of the network, e.g., for the critical distance of two objects above which they are perceived as separate, are derived. The network preferentially chooses the object with the largest angular velocity and the largest angular width. An equation for the velocity and width preferences is presented. Additionally it is shown that for certain neurons of the model, flat receptive fields are more favourable than Gaussian ones. The network exhibits performances similar to those known from amphibians. Various electrophysiological and behavioral results – e.g., the distribution of the diameters of the receptive fields of tectal neurons, of the tongue-projecting salamander Hydromantes italicus and the range of optimal prey velocities for prey catching – can be understood on the basis of the model. Received: 7 December 2000 / Accepted: 13 February 2001