Abstract: In the design and implementation of neural networks, training is determined by a series of architectural and parametric decisions. A method is disclosed that, using genetic algorithms, improves the training characteristics of a neural network. The method begins with a population and iteratively modifies one or more parameters in each generation based on the network with the best training response in the previous generation.

Abstract: The method comprises: selecting an examination window whose size covers the image data associated with a character within a set of characters; presenting image data for a known character to the examining window to obtain a probability density function (PDF) for each pixel within the examining window for each character in the set of characters to be found or segmented to generate a composite PDF for each pixel within the examining window; and using the composite PDF to determine when the examining window is positioned over image data associated with a character within the character set.