There is a growing body of experimental evidence to suggest that the brain is capable of approximating optimal Bayesian inference in the face of noisy input stimuli. Despite this progress, the neural underpinnings of this computation are still poorly understood. In this paper we focus on the problem of Bayesian filtering of stochastic time series. In particular we introduce a novel neural network, derived from a line attractor architecture, whose dynamics map directly onto those of the Kalman Filter in the limit where the prediction error is small. When the prediction error is large we show that the network responds robustly to change-points in a way that is qualitatively compatible with the optimal Bayesian model. The model suggests ways in which probability distributions are encoded in the brain and makes a number of testable experimental predictions.

We introduce a method for the recognition and prediction of motion, based on the idea that different motions trace out different shapes in some state space. In the recognition step we use a multidimensional generalization of the shape context [1] to find the closest prototype motion to the observed data. When tested against motion capture data, our model yields excellent (99%) recognition of gait and good (83%) recognition of identity. In addition to recognition, this process also allows us to find an aligning transform TDP that maps the observed data D onto the prototype P. Given this transform, and its inverse TPD, we use a Bayesian approach to make optimal predictions about the data in the prototype space and then map these predictions back into data space. This approach gives accurate predictions over several gait cycles despite the fact that there is often a significant difference between the observed data and the prototype manifold. 1

Psychophysical studies of contour detection (Kovács et al., IOVS 37: S670 (1996); Nafziger et al., IOVS 39: S859 (1998) indicate that the effective range of surround interactions scale inversely with stimulus density. We present a model based on psychophysical data which suggests that the size of the extra-classical receptive field (CRF) is dynamically modulated by stimulus context. Stimulus density acts to modulate the effective range of long-distance facilitatory connections: high stimulus densities narrow the effective range of the connection field, low densities enlarge the effective surround. Summed facilitatory inputs modulate the gain of the CRF. Simulations match the psychophysical results suggesting that a dynamic receptive field mechanism may underlie the inverse scaling of contour integration with stimulus density.

Recently, we proposed a cortical-based model that accounts for a range of psychophysical results on contour salience [S.C. Yen, L.H. Finkel, Vision Res. 38 (1998) 719–741]. In this model, cortical units are interconnected by long-range horizontal connections where the connection strength between units is defined by an inter-element spatial connectivity function. In the present work, we determine the spatial connectivity function psychophysically and compare the result to the function implemented in the original computational model. The results suggest the original product of gaussian spatial weighting function does not completely describe the psychophysical data. Rather, we find that the connectivity function is “tapered” suggesting a lack of separability between position and orientation in cortical connectivity.

One advantage of cortical synchronization as a binding mechanism is its ability to account for phenomena such as perceptual grouping. Such a mechanism requires the ability to synchronize groups of cells and to desynchronize these groups from each other. We present a striate cortical model of perceptual grouping in which synchronization and desynchronization is carried out by a single, common mechanism. Cortical pyramidal cells and interneurons are simulated using multi-compartmental models. Cells in different orientation columns are inter-connected via two sets of long-distance connections that differ in axonal delays and spatial projections. The relative influence of these two connections determines whether synchronization or desynchronization occurs. Once one group of cells synchronizes, inputs from these cells facilitate synchronization in other orientation columns. We address the role of these synchronizing and desynchronizing connections in mediating perceptual grouping and metastable percepts.

We present a model of contour extraction in which the perceptual salience of contours arises from long-range interactions between orientation-selective filters. It has been previously shown that salient contours may be extracted from noisy images by using a number heuristic features. Our algorithm is based on cortical mechanisms, and simulations show close agreement with results from recent anatomical, physiological and psychophysical studies including recent results of D.J. Field et al. (1993), I. Kovacs et al., and M.K. Kapadia et al. (1995). The performance of the algorithm is demonstrated on a range of psychophysical stimuli and real images.

A biologically-based multistage neural network is presented which produces color constant responses to a variety of color stimuli. The network takes advantage of several mechanisms in the human visual system, including retinal adaptation, spectral opponency, and spectrally-specific long-range inhibition. This last stage is a novel mechanism based on cells which have been described in cortical area V4. All stages include nonlinear response functions. The model emulates human performance in several psychophysical paradigms designed to test color constancy and color induction. We measured the amount of constancy achieved with both natural and artificial simulated illuminants, using homogeneous grey backgrounds and more complex backgrounds, such as Mondrians. On average, the model performs as well or better than the average human color constancy performance under similar conditions. The network simulation also displays color induction and assimilation behavior consistent with human perceptual data.

We propose an algorithm for the determination of three dimensional shape and perspective based on the response of the human visual system to changes in visual textures. Current computer vision algorithms are computationally intensive and show inherent difficulties in integrating additional cues for the determination of shape, such as shading, contour, or motion. In order to develop a fast and simple mechanism less constrained for integrating other cues, we incorporated aspects of the physiological properties of cortical cells in VI into a network model. We provide psychophysical evidence that the local spatial frequency spectrum is represented by the spatially averaged peak frequency (APF). After normalization, this APF measures texture compression and leads to estimates of 3D shape and depth. Simulations of the model show good agreement with human responses to a range of textured images.>

The problems of object segmentation and binding are addressed within a biologically based network model capable of determining depth from occlusion. In particular, the authors discuss two subprocesses most relevant to segmentation and binding: contour binding and figure direction. They propose that these two subprocesses have intrinsic constraints that allow several underdetermined problems in occlusion processing and object segmentation to be uniquely solved. Simulations that demonstrate the role these subprocesses play in discriminating objects and stratifying them in depth are reported. The network is tested on illusory stimuli, with the network's response indicating the existence of robust psychological properties in the system.>

We present a model of how objects can be visually discriminated based on the extraction of depth-from-occlusion. Object discrimination requires consideration of both the binding problem and the problem of segmentation. We propose that the visual system binds contours and surfaces by identifying “proto-objects”—compact regions bounded by contours. Proto-objects can then be linked into larger structures. The model is simulated by a system of interconnected neural networks. The networks have biologically motivated architectures and utilize a distributed representation of depth. We present simulations that demonstrate three robust psychophysical properties of the system. The networks are able to stratify multiple occluding objects in a complex scene into separate depth planes. They bind the contours and surfaces of occluded objects (for example, if a tree branch partially occludes the moon, the two "half-moons" are bound into a single object). Finally, the model accounts for human perceptions of illusory contour stimuli.