Segmentation of touching characters in printed Korean/English document recognition

We present an algorithm for efficient segmentation of the touching characters in printed Korean and English document recognition. We derived two rules to segment touching characters in the bilingual document, one from the shape differences in writing blocks defined in this paper between Korean and English characters, and the other from the reliability factor values generated by the classifiers. The proposed method significantly improves the ability of segmentation and recognition of the actual mixed Korean and English documents