By Topic

Proceedings of 13th International Conference on Pattern Recognition

25-29 Aug. 1996

Go

Filter Results

Displaying Results 1 - 25 of 149
  • Proceedings of the 13th International Conference on Pattern Recognition Volume IV, Track D: Parallel and Connectionist Systems

    Publication Year: 1996
    Request permission for commercial reuse | PDF file iconPDF (485 KB)
    Freely Available from IEEE
  • Autoassociative learning in relaxation labeling networks

    Publication Year: 1996, Page(s):105 - 110 vol.4
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (628 KB)

    We address the problem of training relaxation labeling processes, a popular class of parallel iterative procedures widely employed in pattern recognition and computer vision. The approach discussed here is based on a theory of consistency developed by Hummel and Zucker (1983) and contrasts with a previously introduced learning strategy which can be regarded as heteroassociative, i.e. what is actua... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Incremental distributed classifier building

    Publication Year: 1996, Page(s):128 - 132 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (556 KB)

    In this paper we present a scheme of classification based on a particular processing element (neuron) called yprel. The main characteristics of the approach are: (1) an yprel classifier is a set of yprels networks, each network being associated with a particular class; (2) the learning is supervised and conducted class by class; (3) the structure of the network is not a priori chosen, but is deter... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Author index

    Publication Year: 1996
    Request permission for commercial reuse | PDF file iconPDF (1016 KB)
    Freely Available from IEEE
  • Bayesian adaptation of hidden layers in Boolean feedforward neural networks

    Publication Year: 1996, Page(s):229 - 233 vol.4
    Cited by:  Papers (3)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (444 KB)

    In this paper a statistical point of view of feedforward neural networks is presented. The hidden layer of a multilayer perceptron neural network is identified of representing the mapping of random vectors. Utilizing hard limiter activation functions, the second and all further layers of the multilayer perceptron, including the output layer represent the mapping of a Boolean function. Boolean type... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Piecewise-linear classifiers, formal neurons and separability of the learning sets

    Publication Year: 1996, Page(s):224 - 228 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (380 KB)

    The design of piecewise-linear classifiers from formal neurons is considered. The design classifiers are based on hierarchical, multilayer neural networks. The described procedure allows to find both the structure of network (the numbers of layers and neurons) and weights of single neurons. The main principle of the synthesis procedure is to preserve separability of learning sets during data compr... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recognition of unconstrained handwritten numerals by doubly self-organizing neural network

    Publication Year: 1996, Page(s):426 - 430 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (392 KB)

    In this paper we present an efficient pattern recognizer based on a self-organizing neural network which can adapt its structure as well as its weights. The network, called doubly self-organizing neural network (DSNN), makes use of the structure-adaptation capability to place the nodes of prototype vectors into the pattern space accurately so as to make the decision boundaries as close to the clas... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An RBF network with tunable function shape

    Publication Year: 1996, Page(s):645 - 649 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (352 KB)

    In this paper we consider a radial basis function (RBF) network with tunable shape and spreadout parameters of the activation function. We argue that fewer hidden nodes with different RBF shapes can better match the classification regions still preserving the context of the probabilistic semiparametric approximation of the conditional probability density functions (pdf). Instead of squared Euclide... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • GA-based pattern classification: theoretical and experimental studies

    Publication Year: 1996, Page(s):758 - 762 vol.4
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (372 KB)

    Merits of genetic algorithms (GAs), an efficient evolutionary searching paradigm, are utilized for pattern classification in ℜN by fitting hyperplanes to model the decision boundaries in the feature space. Theoretical analysis establishes that as the size of the training set (n) goes towards infinity, the error probability and the decision boundary of the GA based classifier will ap... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Constructive and robust combination of perceptrons

    Publication Year: 1996, Page(s):195 - 199 vol.4
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (380 KB)

    We propose a new strategy for a constructive training of feedforward neural networks to classify linearly nonseparable patterns. The algorithm results in a configuration of the first layer of the network, which is able to give a faithful internal representation of the input patterns. The weights of the network are obtained by the CadaTron algorithm introduced, which is able to separate clusters of... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiple experts recognition system based on neural network

    Publication Year: 1996, Page(s):452 - 456 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (340 KB)

    For numeral recognition, when a single classifier cannot provide a decision which is 100 percent correct, multiple classifier should be able to achieve higher accuracy. This is because group decisions are generally better than any individual's. In this paper, as evidence, the differences between a ANN classifier and a traditional classifier are discussed. Based on this concept combination methods ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Feedforward neural networks for Bayes-optimal classification: investigations into the influence of the composition of the training set on the cost function

    Publication Year: 1996, Page(s):219 - 223 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (340 KB)

    Under idealized assumptions (infinitely large training sets, ideal training algorithms that avoid local minima and sufficient neural network (NN) structures) trained NNs realize Bayes-optimal classifiers (BOCs) with identical costs as long as the training set is representative. Training sets with relative class frequencies different from the a priori class probabilities implement nonidentical cost... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiresolution skeletonization in binary pyramids

    Publication Year: 1996, Page(s):570 - 574 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (428 KB)

    A multiresolution skeletonization algorithm is presented. The underlying data structure is a binary AND-pyramid. Notwithstanding the fact that the AND-pyramid is likely to break pattern connectedness at lower resolution levels, the algorithm generates connected skeletons at all resolution levels. It can most conveniently be implemented on a parallel SIMD architecture, but efficient MIMD and sequen... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Large-scale parallel data clustering

    Publication Year: 1996, Page(s):488 - 493 vol.4
    Cited by:  Papers (6)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (616 KB)

    Algorithmic enhancements are described that allow large reduction (for some data sets, over 95 percent) in the number of floating point operations in mean square error data clustering. These improvements are incorporated into a parallel data clustering tool, P-CLUSTER, developed in an earlier study. Experiments on segmenting standard texture images show that the proposed enhancements enable cluste... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On-line handwriting character string separation method using network expression

    Publication Year: 1996, Page(s):422 - 425 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (316 KB)

    We propose an on-line handwriting character string separation method which uniformly deals with the character separation features. We devised the string separation method using a multi-level network expression and a method using a unified network expression. The separation performance was improved by unifying the physical feature and logical feature using the network expression. Furthermore, we co... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An efficient method to construct a radial basis function neural network classifier and its application to unconstrained handwritten digit recognition

    Publication Year: 1996, Page(s):640 - 644 vol.4
    Cited by:  Papers (7)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (408 KB)

    This paper describes a method to construct an RBFN classifier efficiently and effectively. The method determines the middle layer neurons by a fast clustering algorithm, APC-III and computes the optimal weights between the middle and the output layers statistically. The proposed method was applied to an unconstrained handwritten digit recognition. The experiment showed that the method could constr... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An improved backpropagation neural network learning

    Publication Year: 1996, Page(s):586 - 588 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (268 KB)

    The backpropagation neural network (BPNN) is a well known and widely used mathematical model for pattern recognition, nonlinear function approximation, time series prediction, etc. There are many applications which require large input and hidden layers. In such cases, the learning process takes a long time. Many authors propose different methods to reduce the learning time, through convergence imp... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Learning class regions by the union of ellipsoids

    Publication Year: 1996, Page(s):750 - 757 vol.4
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (584 KB)

    In many classification schemes objects are represented as points in multi-dimensional feature spaces. The classification scheme then attempts to discriminate between regions in the space occupied by objects of different classes. The performance of the classification method often depends on the shape of the class regions, e.g., whether or not they are linearly separable. In many practical cases, cl... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A neural network-based algorithm to detect dominant points from the chain-code of a contour

    Publication Year: 1996, Page(s):325 - 329 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (436 KB)

    A new algorithm for dominant point detection in chain-coded contours is presented. The algorithm directly operates on the chain-code link values. No computation of the (x,y) co-ordinates of the contour points is done, nor any classical computation of the curvature or its derivative. Instead, a dynamic neural network traverses the contour giving a measurement of the relevance of each point, further... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A computational model of depth-based attention

    Publication Year: 1996, Page(s):734 - 739 vol.4
    Cited by:  Papers (29)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (964 KB)

    We present a computational model for attention. It consists of an early parallel stage with preattentive cues followed by a later serial stage, where the cues are integrated. We base the model on disparity image flow and motion. As one of the several possibilities we choose a depth-based criterion to integrate these cues, in such a way that the attention is maintained to the closest moving object.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Cubical singular simplex model for 3D objects and fast computation of homology groups

    Publication Year: 1996, Page(s):190 - 194 vol.4
    Cited by:  Papers (7)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (352 KB)

    This paper proposes a new simplex model for 3D objects: a cubical singular simplex model instead of the traditional triangularization simplex model. An extended Kohonen mapping is then presented as an efficient learning rule of the model. Based on this model, one can derive a pyramid structure for multi-resolution hierarchy, which is not possible for triangularization simplex. Besides, a fast algo... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Highly accurate recognition of printed Korean characters through an improved grapheme recognition method

    Publication Year: 1996, Page(s):447 - 451 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (420 KB)

    This paper presents a recognition system which obtains a recognition rate higher than 99% for the printed Korean characters of multifont and multisize. The system consists of 18 neural networks: one for character type classifier and the rest for grapheme recognizers. We recognize a given input by first identifying the character type of the input and then recognizing its constituent graphemes. The ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Binary PACT

    Publication Year: 1996, Page(s):606 - 610 vol.4
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (356 KB)

    The pyramid architecture classification tree (PACT) is a novel pattern recognition algorithm which has good capabilities. PACT is a kind of decision tree classifier, though the algorithm is motivated from quite different backgrounds from conventional pattern recognition algorithms. Moreover, PACT motivates us to propose a new hypothesis “a decision region in the feature space having fractal ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Wavelet-FILVQ classifier for speech analysis

    Publication Year: 1996, Page(s):214 - 218 vol.4
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (404 KB)

    This paper describes a novel speech signal classification scheme based on spectrograms which are subjected to wavelet transform: a procedure which yields specific information regarding time and frequency variation of the signal. Feature vectors are extracted and classified using LVQ networks. The output of the network is interpreted as a fuzzy membership coefficient. This scheme is applied to the ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • 3-D object recognition using adaptive scale MEGI

    Publication Year: 1996, Page(s):117 - 122 vol.4
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (476 KB)

    We propose a method for recognition of a 3-D object using a multi scale description of the object and adaptive matching. MEGI model is a description model to represent arbitrary shapes. However, many MEGI elements are necessary to represent uneven or carved surfaces with accuracy, so it is difficult to use them for recognition. As a solution, we make a tree which corresponds to the multi scale des... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.