A Gibbsian Kohonen Network for Online Arabic Character Recognition

The purpose of this study is to investigate handwritten online character recognition by Kohonen neural networks which learn class conditional Gibbs densities from training samples. The characters are represented by histograms (empirical distributions) of features. The Kohonen network learning algorithm implements a gradient ascent which maximizes air entropy criterion under constraints. Using a database of handwritten online Arabic characters produced without constraints by a large number of writers, we conducted extensive experiments which show the advantage of this Cibbsian Kohonen network over other classifiers such as a, regular Kohonen neural network and a Cibbsian Bayes classifier.