Neurofuzzy algorithms have been extensively developed in recent years for the real time/online identification of nonlinear a priori unknown dynamical processes. As with all rule base paradigms they suffer from the cur...
详细信息
Neurofuzzy algorithms have been extensively developed in recent years for the real time/online identification of nonlinear a priori unknown dynamical processes. As with all rule base paradigms they suffer from the curse of dimensionality, restricting their practical use to low dimensional control problems. This paper shows how adaptive construction algorithms based on additive decomposition techniques can overcome this problem, to produce parsimonious neurofuzzy models which retain their transparency or interpretability. Not only does this approach extend the applicability of neurofuzzy algorithms, it also enables low complexity controllers and estimators to be derived. In this context neurofuzzy state estimators are derived, which automatically parameterise a Kalman filter for a process state estimate reconstruction from any input/output data source. This approach avoids pitfalls of the extended Kalman filter, and is optimal for local models. The paper discusses real world applications of this new theory of modelling and estimation to helicopter guidance, intelligent driver warning system, communication antennas, autonomous underwater vehicles, ship collision avoidance guidance, and an IFAC benchmark problem.
This paper presents and evaluates clustering techniques for the training of Radial Basis Function neuralnetworks Moody and Dark (1), Broomhead and Lowe (2). The clustering techniques define the centers of the radial ...
详细信息
This paper presents and evaluates clustering techniques for the training of Radial Basis Function neuralnetworks Moody and Dark (1), Broomhead and Lowe (2). The clustering techniques define the centers of the radial basis functions used by these networks. Therefore, the main purpose is to verify the influence of different clustering techniques in the performance of RBF networks. The K-Means MacQueen (4), widely used for the centers choice in RBF networks, is contrasted with others clustering techniques, such as, Optimal Adaptive K-Means Chinrungrueng and Sequin (5), DHB Duda and Hart (9), DHF Ismail et al (8), AFB Ismail and Kamel (6) and ABF Ismail and Kamel (6). The authors of these techniques claim that they are more likely to converge to an optimal or near-optimal configuration. Initially, the algorithms and a complete description of each technique are presented. Finally, using these techniques the RBF performance in a pattern recognition task is evaluated.
Texture classification and segmentation in digital images is commonly achieved using spatial grey level dependence matrices (SGLDMs), often referred to as co-occurrence matrices. This involves the computation of many ...
详细信息
Texture classification and segmentation in digital images is commonly achieved using spatial grey level dependence matrices (SGLDMs), often referred to as co-occurrence matrices. This involves the computation of many matrices over a range of different spatial separations and orientations. The approach proposed in this paper uses a hybrid neural network system, consisting of a self-organizing map followed by a backpropagation network, to restrict the number of SGLDMs that need to be computed. The system is trained in two phases on images with known texture content. The trained system is able to provide information, in the form of pixel spacing and orientation, on the texture content of unseen images. This information may be used to select appropriate SGLDMs for further texture classification. Experimental results are presented which demonstrate the effective performance of the system.
This paper explores the notion of intelligent components within a supportive architecture, and looks at their applications to multimedia systems. One particular component is considered that is inspired by neural model...
详细信息
This paper explores the notion of intelligent components within a supportive architecture, and looks at their applications to multimedia systems. One particular component is considered that is inspired by neural models;it offers local processing that engenders complex emergent behaviour, providing a simple to use yet powerful visualization tool that has been successfully applied to complex user interaction tasks with a digital library system.
This paper presents a cellular neural network software simulator, called SIMUL CN2, for image-processingapplications. The software is designed to handle with both black-and-white and 256 gray levels images. All templ...
详细信息
ISBN:
(纸本)0780348672
This paper presents a cellular neural network software simulator, called SIMUL CN2, for image-processingapplications. The software is designed to handle with both black-and-white and 256 gray levels images. All template matrices used, are supposed space-invariant with dimension 3x3 or 5x5. The software simulator acts as a development system and an evaluation tool for VLSI chips, currently under study. An automatic optimization tool together with some experimental applications are reported as well.
In this work we present a new proposal for image segmentation using deformable models, as an application of Discrete-Time Cellular neuralnetworks (DTCNN) [1]. This approach is based on active contours (also called sn...
详细信息
ISBN:
(纸本)0780348672
In this work we present a new proposal for image segmentation using deformable models, as an application of Discrete-Time Cellular neuralnetworks (DTCNN) [1]. This approach is based on active contours (also called snakes) which evolve until reaching a final desired location. The contours are guided by both external information from the image under consideration which attracts them rewards salient characteristics of the scene, and internal energy from the contour image which tries to maintain the smoothness in the curve shape. The massively parallel processing in DTCNN and the use of local information permit a VLSI implementation suitable for real time applications.
The three-dimensional reconstruction of human body parts, and faces in particular, is catalyzing growing interest in many disciplines ranging from basic imageprocessing to video conferencing, constructive and plastic...
详细信息
The three-dimensional reconstruction of human body parts, and faces in particular, is catalyzing growing interest in many disciplines ranging from basic imageprocessing to video conferencing, constructive and plastic surgery, rehabilitation and virtual clones. Two main problems which have to be faced: filtering of the noise associated to sampling and interpolation between the samples. These two problems can be reframed in the domain of regularization. It is shown that a regularized model can be efficiently obtained by using a new neural network called hierarchical radial basis functions network (HRBF).
In this paper an optoelectronic model of discrete rime cellular neuralnetworks (DTCNN) is presented. Connections between cells and parallel input-output ar-e realized using optoelectronic devices. As an emitter-recei...
详细信息
ISBN:
(纸本)0780348672
In this paper an optoelectronic model of discrete rime cellular neuralnetworks (DTCNN) is presented. Connections between cells and parallel input-output ar-e realized using optoelectronic devices. As an emitter-receiver device an optical thyristor is applied. Connections between cells are realized by diffractive Damman gratings. We propose a dual rail system for early processing of binary images. The CNN system performs mathematical morphology and space logic operations.
In this paper Lyapunov Diagonally Stable matrices are used to design Cellular neuralnetworks for associative memories. The proposed technique, which guarantees the global asymptotic stability of the equilibrium point...
详细信息
ISBN:
(纸本)0780348672
In this paper Lyapunov Diagonally Stable matrices are used to design Cellular neuralnetworks for associative memories. The proposed technique, which guarantees the global asymptotic stability of the equilibrium point, generates neural circuits where the input data are fed via external inputs, rather than initial conditions. This feature makes the suggested approach particularly suitable for hardware implementation techniques. Simulations results are reported to show the advantages and the usefulness of the proposed design method.
Insight into the core of the Pipelined Recurrent neural Network (PRNN) in prediction applications is provided. It is shown that modules of the PRNN contribute to the final predicted value at the output of the PRNN in ...
详细信息
Insight into the core of the Pipelined Recurrent neural Network (PRNN) in prediction applications is provided. It is shown that modules of the PRNN contribute to the final predicted value at the output of the PRNN in two ways, namely through the process of nesting, and through the process of learning. A measure of the influence of the output of a distant module to the amplitude at the output of the PRNN is analytically found, and the upper bound for it is derived. Furthermore, an analysis of the influence of the forgetting factor in the cost function of the PRNN to the process of learning is undertaken, and it is found that for the PRNN, the forgetting factor can even exceed unity in order to obtain the best predictor. Simulations on three speech signals support that approach, and outperform the other stochastic gradient based schemes.
暂无评论