Projective distortion of surface texture observed in a perspective image can provide direct information about the shape of the underlying surface. Previous theories have generally concerned planar surfaces;this paper ...
详细信息
Projective distortion of surface texture observed in a perspective image can provide direct information about the shape of the underlying surface. Previous theories have generally concerned planar surfaces;in this pap...
详细信息
In this paper we study small depth circuits that contain threshold gates (with or without weights) and parity gates. All circuits we consider are of polynomial size. We prove several results which complete the work on...
详细信息
A highly simplified network model of cortical associative memory, based on Hebb's theory of cell assemblies, has been developed and simulated. The network comprises realistically modelled pyramidal-type cells and ...
A highly simplified network model of cortical associative memory, based on Hebb's theory of cell assemblies, has been developed and simulated. The network comprises realistically modelled pyramidal-type cells and inhibitory fast-spiking interneurons and its connectivity is adopted from a trained recurrent artificial neural network. After-activity, pattern completion and competition between cell assemblies is readily produced. If, instead of pyramidal cells, motor neuron type cells are used, network behaviour changes drastically. For instance, spike synchronization can be observed but after-activity is hard to produce. The authors results support the biological feasibility of Hebb's cell assembly theory. The analogy between this theory and recurrent artificial neural network models is discussed.
The paper outlines a method for designing near optimal nonlinear classifiers based on a self-organizing technique for estimating probability density functions when only weak assumptions are made about the densities. T...
详细信息
The paper outlines a method for designing near optimal nonlinear classifiers based on a self-organizing technique for estimating probability density functions when only weak assumptions are made about the densities. The classical parametric and nonparametric methods for estimating density functions have a number of drawbacks;parametric methods give weak results on unknown distributions, while nonparametric methods require extensive amounts of design samples, storage capacity, and computing power. The present method avoids these disadvantages by parameterizing a set of component densities from which the actual densities are constructed. The parameters of the component densities are optimized by a self-organizing algorithm, reducing to a minimum the labeling of design samples. All the required computations are realized with the simple "sum of product" units commonly used in connectionist models. The density approximations produced by the method are illustrated in two dimensions for a multispectral image classification task. The practical use of the method is illustrated by a small speech recognition problem, that of recognizing 18 Swedish consonants. Related issues of invariant projections, cross-class pooling of data, and subspace partitioning are also discussed.
We investigate the power of threshold circuits of small depth. In particular, we give functions that require exponential size unweighted threshold circuits of depth 3 when we restrict the bottom fanin. We also prove t...
详细信息
We give a probabilistic algorithm for computing the greatest common divisor (GCD) of two polynomials over an algebraic number field. We can compute the GCD using O(llog5(l)) expected binary operations where l is the s...
详细信息
This report describes a three-dimensional computer model of the olfactory cortex developed for the study of cortical oscillations and their biological significance. The model was designed with the intention of investi...
This report describes a three-dimensional computer model of the olfactory cortex developed for the study of cortical oscillations and their biological significance. The model was designed with the intention of investigating the relative role of network circuitry and network unit properties, resulting in a model complexity between simple Hopfield nets and detailed realistic simulations. Network connections are essentially the same as in a detailed simulation of the olfactory (piriform) cortex by Wilson and Bower (1989), but the network units are here modeled with continuous output functions and single compartments. It is shown that the present model is capable of reproducing all major results of the more complex model, corresponding to spatiotemporal patterns found in the actual cortex (Freeman 1975). This indicates that action potentials and the geometry of cells are not needed per se for explaining certain cortical activities. In contrast, connections between units, in particular feedforward and feedback inhibitory loops and long-range, excitatory-excitatory connections, are found to be crucial for the dynamical behavior of this system. The model describes intrinsic oscillatory properties of olfactory cortex and reproduces response patterns associated with a continuous random-input signal and with a shock pulse given to the cortex. In the latter case, waves of activity move across the model cortex in a way similar to the detailed simulations by Wilson and Bower, and consistent with corresponding global dynamic behavior of the functioning cortex. For a constant random input, the network is able to oscillate with two separate frequencies simultaneously, purely as a result of its intrinsic network properties. A delicate balance between inhibition and excitation, in terms of connection strength and timing of events, is necessary for coherent frequency and phase of the oscillating neural units. The analytical equations used in this model seem an adequate representation
作者：
Lars LangemyrNumerical Analysis and Computing Science
Royal Institute of Technology S-100 44 Stockholm Sweden and Wilhelm-Schickard-Institut für Informatik Universitat Tübingen Auf dem Sand 13 W-7400 Tübingen Germany
Focus-of-attention is extremely important in human visual perception. If computer vision systems are to perform tasks in a complex, dynamic world they will have to be able to control processing in a way that is analog...
详细信息
Focus-of-attention is extremely important in human visual perception. If computer vision systems are to perform tasks in a complex, dynamic world they will have to be able to control processing in a way that is analogous to visual attention in humans. Problems connected to foveation (examination of selected regions of the world at high resolution) are examined. In particular, the problem of finding and classifying junctions from this aspect is considered. It is shown that foveation as simulated by controlled, active zooming in conjunction with scale-space techniques allows for robust detection and classification of junctions.
暂无评论