Hypergeometric Laguerre moment for handwritten digit recognition

This paper introduces a new set of orthogonal moment functions based on the discrete Laguerre polynomials to evaluate a set of candidate features and to select an informative subset to be used as input data for a neural network classifier.The implementation of moments proposed in this paper does not involve any numerical approximation. The first step (pre-processing) of the proposed method takes into account the discriminative properties of Laguerre moments and proposes a novel method that extracts optimal object features. for this, we introduce the Maximum Entropy Principle (MEP) as a selection criterion. The second step (recognition) is achieved by using Multilayer Feedforward Neural Network (MFNN) as a classifier with the stochastic back propagation as a learning algorithm. Finite vectors obtained as a result in the pre-processing phase are then fed into the neural network system. The proposed method has been tested on the well known MNIST database of handwritten digits. It produces excellent and encouraging results by reducing the computational burden of the recognition system and presenting a high recognition rate with good generalization ability.