A unified network-based approach for recognition of cursive handwritings in mixed languages: a case study on Hangul and Roman mixture

Recent development of pen computing technology enables handwriting to be one of the major modes of human and computer interaction. Although handwriting recognition capability is not sufficiently acceptable for practical use, several systems recognizing Roman characters are available in the market. On the other side of the globe, non-Roman character recognition systems are also under development to respond the pen computing technology. Japanese Kana and Korean Hangul recognition technologies are seemingly reaching a stage of practicality. The authors propose a unified network based approach for recognizing freely handwritten text in more than one language. This approach can be applied to any combination of phonetic writing systems including Arabic, Tai and Japanese. Although intensive evaluation yet to come, initial implementation for Hangul and Roman mixture with digits yields promising results which cannot be easily surpassed by other approaches. By combining component languages, recognition accuracy drops negligibly little, but speed is slowed substantially. More powerful search methods and machines are needed to use such an approach in practice