A new algorithm of mixed Chinese-English character segmentation based on irregularty degree

The character segmentation, as an important part of the character recognition technology, directly restricts the development of the character recognition. In this paper, after analyzing some traditional algorithms for character segmentation, a new algorithm of segmentation based on irregularity degree is proposed which can first label the characters and then segment them using different ways. The experimental results demonstrate the applicability and favorable performance of the proposed algorithm.