Off-line recognition method for handwritten Chinese character based on mathematical morphology
基于数学形态学的脱机手写体汉字识别方法

The stroke feature is efficient for off-line handwritten Chinese character recognition. But the process for thinning strokes always makes the characters distorted. A stroke feature extraction approach that doesn't need the thinning preprocess was proposed. By erosion and dilation operations of the mathematical morphology and using different adaptive structure elements, the characters were decomposed into strokes. Then the elastic meshes were applied to extract the directional feature. Finally, generalized K-L transform was used to decrease the dimensions of the feature vectors so as to get rid of redundancies. Experiment results show the effectiveness of the proposed approach.