MEDIAN RADIAL BASIS FUNCTION NEURAL-NETWORK

Radial basis functions (RBF's) consist of a two-layer neural network, where each hidden unit implements a kernel function, Each kernel is as sociated with an activation region from the input space and its output is fed to an output unit, In order to find the parameters of a neural network which embeds this structure we take into consideration two di fferent statistical approaches, The first approach uses classical esti mation in the learning stage and it is based on the learning vector qu antization algorithm and its second-order statistics extension, After the presentation of this approach,we introduce the median radial basis function (MRBF) algorithm based on robust estimation of the hidden un it parameters. The proposed algorithm employs the marginal median for kernel location estimation and the median of the absolute deviations f or the scale parameter estimation, A histogram-based fast implementati on is provided for the MRBF algorithm, The theoretical performance of the two training algorithms is comparatively evaluated when estimating the network weights, The network is applied in pattern classification problems and in optical flow segmentation.