HomeChair: Olli Yli-Harja, Helsinki University of Technology, Finland
Home
Adrian G. Bors, University of Thessaloniki (Greece)
Ioannis Pitas, University of Thessaloniki (Greece)
In this paper we use Radial Basis Function (RBF) networks for object modeling in images. An object is composed from a set of overlapping ellipsoids and has assigned an output unit in the RBF network. Each basis function can be geometrically represented by an ellipsoid. We introduce a new robust statistics based algorithm for training radial basis function networks. This algorithm relies on alpha-trimmed mean statistics. The use of the proposed algorithm in estimating ellipse parameters is analyzed.
ns970211.pdf (Scanned) ns970211.pdf (From Postscript)
TOP
Tadasuke Inoue, Musashi Institute of Technology (Japan)
Akira Taguchi, Musashi Institute of Technology (Japan)
Nonlinear filters which are utilized rank-order information and temporal-order information, have many proposed, in order to restore nonstationary signals which are corrupted by additive noise. LWOS (Linear Combination of Weighted Order Statistics) filters [11 which also utilized two informations , and have properties of efficient impulsive and nonimpulsive noise attenuation and sufficiently details and edges preservation. In this paper, we propose a data-dependent LWOS filter whose coefficients change based on local statistics.
ns970212.pdf (Scanned) ns970212.pdf (From Postscript) TOPOlli Yli-Harja, University of Helsinki (Finland)
In this paper the properties of the joint distribution function of the outputs of stack filters with common arguments are examined. The special characteristics of these functions are discussed. Different approaches for characterising them are considered. Empirical tests are performed and reported. Different ways of extracting useful correlation information are compared.
ns970213.pdf (Scanned) ns970213.pdf (From Postscript) TOP
Akira Taguchi, Musashi Institute of Technology (Japan)
Susumu Takaku, Musashi Institute of Technology (Japan)
Stack filters are a class on nonlinear filters and include rank order filters, morphological filters, weighted median filters, and so on. The stack filter is defined by a Boolean function. The output of Boolean function is restricted two values (i.e., 0 or 1). We attempt to enlarge class of stack filter by defining the output of Boolean function from 0 to 1 continuously. We call this filter the fuzzy stack filter. We have already proposed fuzzy weighted median (FWM) filters which are important class of fuzzy stack filters, and shown a simple design method of those. However, this design method imposed restrictions on the class of FWM filter which is able to be design. In this paper, we propose a novel design method of FWM filters by using LMS algorithm. We can design the optimal FWM filters form all class of those by the proposed method.
ns970214.pdf (Scanned) ns970214.pdf (From Postscript) TOP
Vassilios Chatzis, University of Thessaloniki (Greece)
Ioannis Pitas, University of Thessaloniki (Greece)
In this paper, the Fuzzy Vector Median is proposed, defined as an extension of Vector Median. It is based on a novel distance definition of multidimensional fuzzy numbers (fuzzy vectors), which satisfy the property of angle decomposition. The proposed distance of two fuzzy vectors depends on the classical distance of the fuzzy set centers and on the fuzziness that every fuzzy set holds. As a result the Fuzzy Vector Median of a set of fuzzy vectors is affected by the presence of fuzziness.
ns970215.pdf (Scanned) ns970215.pdf (From Postscript) TOPAleksej Markarov, Swiss Federal Institute of Technology (EPFL) (Switzerland)
A cyclic component signal is a nonstationary signal defined as superposition of a trend, of one or more almost periodicities and an uncorrelated stochastic component. The cycles of almost-periodicities are not, in contrast to periodicities, the shifted replicas of each other, but vary over time in wavelength, amplitude and shape. The almost-periodic components of a signal are often referred to as cyclic components 111
ns970216.pdf (Scanned) ns970216.pdf (From Postscript) TOP