A practical rough classifier for Chinese handwriting recognition

It is very practical to apply multilevel classifier on solving large scale and complex recognition problems such as Chinese handwriting recognition. A multilevel classifier includes a high-level classifier (rough classifier) whose responsibility is to split patterns into different groups. In this paper, we design a novel rough classifier system that is used for Chinese handwriting recognition. Our algorithm is composed of two parts: clustering and classify. Our experiments show that the system achieves high precision on both training data and testing data.