Handwritten Chinese character recognition with directional decomposition cellular features

A new feature extraction approach based on elastic meshing and directional decomposition techniques for handwritten Chinese character recognition (HCC R) is proposed in this letter. It is found that decomposing a Chinese chara cter into horizontal, vertical stroke, left slant and right slant direction al sub-patterns is very helpful for feature extraction and recognition. Thr ee kinds of decomposition methods are proposed. A minimum distance classifi er is trained by 3755 categories of characters using the new features. Test ing on a total of 37,550 untrained handwritten samples produces the recogni tion rate of 92.36%, showing the effectiveness of the proposed approach.