Multiresolution recognition of offline handwritten Chinese characters with wavelet transform

The authors propose a novel multiresolution recognition scheme for handwritten Chinese character recognition in which an input pattern is recognized by adopting the coefficients of the wavelet transforms. It is known that wavelet representation provides a coarse-to-fine strategy. The recognition starts from the coarse scale and moves to the finer scales. After preprocessing, the wavelet transform is applied to the kanji image. Then, we make use of the coefficients with the lowest resolution to select 50 candidates from 3755 categories. In order to enhance the statistical feature of a character, we used statistical methods to reconstruct features in fine classification. With the proposed recognition system, experiments are performed on the 863 Testing System. The correct rate reaches 80.56%, which is a promising result.