Rough sets and neural networks application to handwritten character recognition by complex Zernike moments

Abstract

The paper presents a hand-written character recognition by the data mining and knowledge discovery software system RoughNeu- ralLab. In recognition experiments the Zernike moments were applied as the extracted features of character images. For further feature reduction the rough set theory method was applied as a front end of neural network. Eventually the error backpropagation neural network classifiers were designed for the reduced feature subsets.