L.O. Chua and L. Yang. Cellular Neural Networks: Theory and applications. IEEE Trans. Circuits and Syst., 35:1257--1290, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and others) nevertheless, despite the nature of the stimuli and the biological evidence of the neurons operation in 2D and 3D, only a few number of models take advantage of the properties of the multidimensional schemata. Some of these models are: the Kohonen SOM [6] Cellular Neural Networks [7], and the more recently proposed 2D Modified Kanerva s network [5] Despite of the fact that Neurophysiology has not currently greatly influence in motivating artificial connectionist models (ACM) operating in more than one dimension, it has in the matter of the information storage structures, and ....

L.O. Chua & L. Yang, "Cellular Neural Networks: Theory and applications", IEEE Trans. Circuits and Systems, Vol. 35, pp. 1257-1290, 1988.

....a small number of points in time. Our results demonstrate that CNN obtained with our method approximate the dynamical behaviour of various nonlinear systems accurately. Results for two nonlinear PDE, the Phi 4 equation and the sine Gordon equation, are discussed in detail. 1. INTRODUCTION CNN [1,2,3] form a special class of recurrent neural networks with the following distinguishing properties: ffl The cells, the states and outputs of which are given by real numbers, are placed in one ore more layers on a regular lattice usually of 1 , 2 or 3 dimensions. ffl Direct interactions between ....

Chua, L.O. and Yang, L.: Cellular Neural Networks: Theory and Applications. IEEE Transactions on Circuits and Systems, vol. 35, pp. 1257--1290 (1988)

....2000. In this part of the contribution the basic structure and features of SCNN 2000 will be discussed, whereas the SCNN control system is presented in a second paper [2] 1. Introduction Since it s rst presentation in [1] SCNN 1 has become one of the mostly used simulation systems for CNN [3]. It operates under di erent systems, like AIX UNIX, SGI UNIX, HP UNIX, Sun Solaris, Linux and Microsoft Windows. SCNN 2000 has a nearly unlimited capability for precise CNN simulations and has not the limitations of older versions, for example a simulation was restricted to 1 or 2 dimensional ....

Chua, L.O. and Yang, L.: \Cellular Neural Networks: Theory and Applications", IEEE Trans. on Circuits and Systems, Vol. 35. Nr.10, pp. 1257-1290, 1988.

....of deviated parameters is shown for different image processing templates. We propose a new parameter learning method for minimizing the effect of template and bias deviations. In all treated cases a significant improvement can be observed by using this method. 1. Introduction Since Chua et al.[1] have introduced CNN, the hardware realization of the CNN universal machine [2] has been subject in many investigations [3,4,5,6] Espejo et al. 7] presented a CNN VLSI chip with 22 Theta 20 neurons and binary valued outputs, another VLSI prototype with continuous cell outputs and a size of 16 ....

....results as it will be shown later. In our investigations we consider a single layer CNN with the set of state equations dx i (t) dt = Gammax i (t) X j2N(i) A i Gammaj y j (t) B i Gammaj u j I; 1) where depending on the special case, y i (t) defines the piecewise linear output function [1], y i (t) f(x i (t) 1 2 (jx i 1j Gamma jx i Gamma 1j) 2) or the sigmoidal function y i (t) f(x i (t) 2 1 e Gammafix i (t) Gamma 1: 3) In this contribution the influence of hardware tolerances were modeled by changing the network parameters according to A i Gammaj ) A i;j = ....

Chua, L.O. and Yang, L.: Cellular Neural Networks: Theory and Applications; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257--1290 (1988)

....Cellular Neural Networks is presented based on evolutionary strategies. The proposed global optimization procedure is discussed in detail and the performance on various parameter determination problems will be shown afterwards. 1. Introduction The universal Cellular Neural Networks (CNN) paradigm [1] has been studied in various investigations, which are leading to many different applications, e.g. image processing, by solving nonlinear partial differential equations or by modelling complex natural phenomena. Generally a CNN is an arrangement of coupled cells, where all cells interact only ....

L.O. Chua and L. Yang: " Cellular Neural Networks: Theory and Applications"; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257--1290, 1988.

....(PDE) Our results show that depending on the training pattern solutions of various PDE can be approximated with high accuracy by a simple CNN structure. Results for two nonlinear PDE, Burgers equation and the Korteweg de Vries equation, are discussed in detail. 1. INTRODUCTION A CNN [2,3,4] is a system of simple nonlinear processors (cells) which are arranged in one or more layers on a regular grid. Interactions between cells are local and usually translation invariant, i.e. a connection from a cell j towards another cell i only exists if j is part of i s neighborhood N (i) and its ....

Chua, L.O. and Yang, L.: Cellular Neural Networks: Theory and Applications; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257--1290 (1988)

....for networks with translation variant and invariant templates are implemented in SCNN. As an example, parameter deviations of a template have been reduced by training. Simulation and training results will be discussed in detail. 1. Introduction The dynamics of a Cellular Neural Network (CNN) [1,2,3] is determined by state equations of the form x dx m i (t) dt = Gammaffx m i (t) M X m 0 =1 X j2N m 0 m (i) A m 0 m i;j (y m 0 j (t) y m i (t) B m 0 m i;j (u m 0 j (t) u m i (t) 1) A 0 m 0 m i;j (y m i (t Gamma A ) y m 0 j (t Gamma B ) ....

....connections between the cells, accordingly B m i;j are functions of the cell inputs u m j (t) The weights A 0 m i;j and B 0 m i;j are the delaytime templates. I m i (t) is the bias of each cell. 2. SCNN: A universal simulating system for CNN Since CNN have been proposed by Chua and Yang [1], they have found many of important applications in signal and image processing. CNN simulators are necessary tools for developing and testing algorithms running on these networks. Roska et al. presented the first widely used simulation system [7] which allows the simulation of a large class of ....

Chua, L.O. and Yang, L.: Cellular Neural Networks: Theory and Applications; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257--1290 (1988)

....zquez I. INTRODUCTION Cellular Neural Networks (CNNs) are arrays of identical nonlinear dynamic processing units (cells) arranged on a regular grid where direct interactions among cells are limited to a finite local neighborhood. CNNs were first proposed by L.O. Chua and L. Yang in 1988 [1] and have an architecture similar to cellular automata, although differ in that interactions among cells are analog, and in the dynamic nature of the processing performed by the cells. A primary reason for the interest of CNNs is the existence of many computational and signal processing problems ....

....an assembly of synchronized (if required) layers conforms the top level network definition. Let us describe some examples with different CNN implementations to illustrate the modeling capabilities of SIRENA. a) Chua and Yang original CNN model The original model proposed by Chua and Yang [1] describes time evolution and neighborhood coupling of each individual cell (c) within the grid domain (GD) in terms of a network time constant (t) a radius of vicinity (r) feedback and control templates (a c d and b c d ) that weight the influence of the neighbors input and output variables ....

L.O. Chua and L. Yang: "Cellular Neural Networks: Theory and Applications". IEEE Trans. Circuits and Systems, Vol. 35, pp 1257-1290, October 1988.

....and obtaining translation invariant templates and a translation variant bias, which is well suited for hardware implementations with nonideal components. First results obtained with a hardware environment are shown. I. INTRODUCTION The dynamics of a multi layer Cellular Neural Network (CNN) [1,2,3] a system of nonlinear local interacting cells, can be determined for example by state equations of the form C dx m i (t) dt = Gamma 1 R x m i (t) M X m 0 =1 (1) X j2N m 0 m (i) A m 0 m i;j (y m 0 j (t) y m i (t) X j2N m 0 m (i) B m 0 m i;j (u m 0 ....

Chua, L.O. and Yang, L.: Cellular Neural Networks: Theory and Applications; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257-- 1290 (1988)

....to be performed on the whole image. This concerns early vision, but also image segmentation and pattern recognition. Though not incompatible, programmability and analog VLSI do not get al..ong well together as demonstrated for example by the limitations of CNNs (Cellular Non linear Networks) [4]. So PARs are preferably based on digital circuit techniques, yet nonstandard ones due to the severe constraints on silicon area. However, the cohabitation with analog preprocessing circuitry is possible and even desirable as suggested by figure 1. digital PE photosensor analog proc. ....

L.O. Chua and L. Yang. Cellular neural networks: Theory and applications. IEEE Trans. on Circuits and Systems, 35(10):1257-1290, October 1988.

....into account the fact that those operators usually do not use the same operating ranges or the same physical quantities to represent information. The most advanced attempt to bring together analog computation and programmability can be found in structures called Cellular Neural Networks (CNN)[3]. These CNNs are controlled through a set of 19 parameters distributed as voltages within each pixel. However, 6 to 7 bits dynamics must be achieved on these control parameters to ensure robust operation, so fixed pattern noise in the array becomes a major issue. This leads to using the MOS ....

L.O. Chua and L. Yang. Cellular neural networks: Theory and applications. IEEE Trans. on Circuits and Systems, 35(10):1257--1290, October 1988.

....through sequential and bifurcated flow algorithms. This paper presents a 0.8m CMOS vision chip prototype which follows these guidelines. The processing function is based on the paradigm of Cellular Neural Networks (CNN) a very suitable framework for systematic design of image processing chips [3]. The complete programmability of the interconnection strengths, its internal image memories, and other additional features make the prototype a powerful front end for the realization of simple and medium complexity artificial vision tasks. Also, although the fundamental processing function is ....

....a zero mean distribution of pixel values. Optional external circuitry can be employed to adjust the mean of the distribution when needed, for instance for highly regular images with dominant background. Processing: Image processing is based on the Cellular Neural Network (CNN) computing paradigm [3]. CNNs can be described as artificial neural networks (ANNs) with neurons (or cells) spatially distributed on a regular grid (over the chip surface) and locally interconnected. Regularity serves to the purpose of scene discretization into a pixel matrix, while local connectivity, commonly ....

L.O. Chua and L. Yang.: "Cellular Neural Networks: Theory and Applications". IEEE Trans. Circuits and Systems, Vol. 35, pp. 1257-1290, October 1988.

....are special and do not require the 32 bit floating point accuracy. The alternative is the analogic CNN array computer performing about Tera equivalent operations per second, however, on a single chip. Ten years ago, in the seminal, paradigm forming, and now historic paper L.O.Chua and L. Yang [1] introduced the Cellular Neural Network (CNN) now we call also Cellular Nonlinear Network, as a 2D or 3D regular array of locally interconnected nonlinear dynamic systems called neurons, or cells, whose global functionality is determined by a small number of parameters. These parameters define ....

....to determine local template elements, e.g. local illumination is calculated and set to the local bias terms (z) The key issue is that the number of independently adapting values are small; instead of 19, one or two. For example, in a 4 element LAM if we have 2 TCM values (c1, c2) in a LAM: LAM [1 4] : a1 a2 c1 c2 then we may use e.g. z = c1 or A or B = c2 c1 c2 c1 aoo or boo c1 c2 c1 c2 . slowly time varying plasticity is governed by local template control signals (TCS) During, or within, a finite potentiation time Tp, a Potentiation Rule operator (implemented in the LAOU) will determine ....

[Article contains additional citation context not shown here]

L. O. Chua and L. Yang, "Cellular neural networks: Theory and Applications", IEEE Transactions on Circuits and Systems, Vol.35, pp.1257-1290, 1988.

....Department of Electrical Engineering, Katholieke Universiteit Leuven, K. Mercierlaan 94, B 3001 Leuven, Belgium email: FVDSRGL#V]WDNL#KX# ABSTRACT: This paper is concerned with the implementation of cryptographic hash functions on the regular array of simple cellular neural network (CNN, [1, 2]) cells with periodic boundary conditions. Cryptographic hash functions enable message origin authentication and validation of message content integrity. A class of cryptographic hash functions termed Cartesian authentication codes provide provable (unconditional) security for message ....

L.O. Chua and L. Yang, `Cellular neural networks: theory and applications', IEEE Trans. Circuits and Systems, vol.35, no.10, 1257-1290 (1988).

....works may not be reported without the explicit permission of the copyright holder. Abstract This paper presents a systematic approach to design CMOS chips with concurrent picture acquisition and processing capability. Pixel smartness is achieved by exploiting the Cellular Neural Network paradigm [1], incorporating at each Spixel an analog computing cell which interacts with those of nearby Spixels. We propose a current mode technique for CNNSpixel chips and give measurements from two 16 16 prototypes in a single poly double metal CMOS n well 1.6 m technology. One of these prototypes is ....

....a few rows and columns at the grid borders [7] In addition to their usage for preprocessing tasks, smart pixel chips are also useful as standalone units for non intensive computation tasks such as halftoning, motion detection, range finding, etc. The paradigm of Cellular Neural Networks (CNN) [1] is a very suitable framework for systematic design of Spixel chips. On one hand, CNNs consist of regular arrangements of cells topologically identical to Spixel chips. On the other, their cells are only locally connected, and, thus require simple routing. Also, the vast body of literature on ....

[Article contains additional citation context not shown here]

L.O. Chua et al.: "Cellular Neural Networks: Theory and Applications ". IEEE Trans. Circuits and Systems, Vol. 35, pp. 1257-1290, 1988.

....therefore it is suggested to use it as part of a CNN based mammogram analysis system. The paper also gives a CNN based method for the fast implementation of the MPEG and JPEG moving and still image compression standards. 1 Introduction The Cellular Neural Nonlinear Network (CNN) paradigm [1, 2] is in an intimate relationship with image processing. The CNN Universal Machine (CNNUM) 3] architecture provides for a stored program spatiotemporal universal chip and for the innovative application of programmed biological information processing mechanisms. There are four major areas in image ....

L. O. Chua and L. Yang. "Cellular Neural Networks: Theory and Applications". IEEE Trans. on Circuits and Systems, 35:1257--1290, 1988.

....coefficients. Assuming an input image size of N Theta N pixels, the following amount of operations is required: Operation Multiplications Additions 40 Convolutions with 242 coefficients, N Theta N pixels 9680 N 2 9640 N 2 2. 4 Neural fields Also, the computation of neural fields according to [3, 6] and of cellular neural networks by recurrent convolution with arbitrary neighbourhood radius are supported and require the following amount of operations: Operation Multiplications Additions N Theta N Neural Field with 7 Theta 7 weights, one iteration 49 N 2 50 N 2 3 Processor Design In ....

.... to simulation results, the following accuracies were determined for the algorithms: Algorithms Cache Resolution Coefficient Result Resolution Weight Resolution Gaussian Filtering 8 bit 8 bit 8 bit Wavelet transformation 8 bit 16 bit 8 bit Gabor transformation 8 bit 32 bit 32 bit Neural fields [3, 6] 32 bit 32 bit 32 bit We concentrated on an optimization of the processor architecture with respect to the acceleration of convolutions. In our architecture the convolution computation for a pixel position is sequentially carried out by a single processing element, but 64 pixel positions are ....

Chua, L. O. and Yang, L.: Cellular Neural Networks: Theory and Applications, IEEE Transactions on Circuits and Systems, Vol. 35, No. 10, 1988, pp 1257 -- 1290

....to some other classification methods, e.g. the Hamming distance calculation. A number of tests have been completed within the so called bubble debris segmentation experiments using original and artificial gray scale images. 1. Introduction Since the publication of the original paper in 1988 ([1]) the rapidly growing field of Cellular Neural Networks (CNNs) have found numerous potential applications, especially in image processing problems where real time signal processing is required. The CNN approach evolved to a widely accepted computational paradigm [2] and recently its dedicated ....

L. O. Chua and L. Yang, "Cellular Neural Networks: Theory and Applications", IEEE Trans. on Circuits and Systems, Vol. 35, pp. 1257-1290, Oct. 1988.

No context found.

L.O. Chua and L. Yang. Cellular Neural Networks: Theory and applications. IEEE Trans. Circuits and Syst., 35:1257--1290, 1988.

No context found.

L. O. Chua and L. Yang, "Cellular neural networks: Theory and Applications," IEEE Trans. on Circuits and Syst. I, vol.35, pp. 1257-1290, Oct. 1988.

No context found.

L.O.Chua and L.Yang, "Cellular neural networks: Theory and Applications ", IEEE Trans. on Circuits and Systems, Vol.35, pp. 1257-1290, 1988.

No context found.

L.O. Chua and L. Yang: " Cellular Neural Networks: Theory and Applications"; IEEE Transactions on Circuits and Systems vol. 35, pp. 1257--1290 (1988)

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC