A Radical Cascade Classifier for Handwritten Chinese Character Recognition

Radical extraction is the core technique for radical-based Chinese character recognition. In this paper, we proposed a new method of radical extraction – radical cascade classifier. The radical cascade classifier consists of multiple AdaBoost classifiers. It can detect and extract specific radical from characters. To apply cascade classifier to radical extraction, we focus on two main points: feature selection and radical detection. In this paper, we discussed feature selection for the radical cascade classifier and proposed two methods of radical detection. Based on these works, we constructed the radical cascade classifier and conducted experiments on HITPU databases. The experimental results have shown that our approach is efficient.