Nonlinear shape normalization methods for the recognition of large-set handwritten characters

Recently, several nonlinear shape normalization methods have been proposed in order to compensate for shape distortions in large-set handwritten characters. In this paper, these methods are reviewed from the two points of view: feature projection and feature density equalization. The former makes feature projection histogram by projecting a certain feature at each point onto horizontal- or vertical-axis and the latter equalizes the feature densities of input image by re-sampling the feature projection histogram. Then, the results of quantitative evaluation for these methods are presented. These methods have been implemented on a PC in C language and tested with a large variety of handwritten Hangul syllables. A systematic comparison of them has been made based on the following criteria: recognition rate, processing speed, computational complexity and degree of variation.