Learning algorithms for a neural network with laterally inhibited receptive fields

This paper presents a neural network with its output layer as a classifier and its hidden layer constrained by laterally inhibited receptive fields as feature extractor, in which the idea that wavelet transforms are very suitable for modeling the primary visual information processing is reflected. Two learning algorithms for designing the receptive fields are proposed. The problem associated with local minima caused by the inherent oscillatory property in laterally inhibited receptive fields is overcome in the algorithm using discrete wavelets. Good performance is obtained in the experiment of ECG signal classification using the neural network