AN APPROACH TO INTEGRATED PEN INTERFACE FOR JAPANESE TEXT ENTRY

An integrated pen interface system was developed to allow effective Ja panese text entry It consists of subsystems for handwriting recognitio n, contextual post-processing, and enhanced Kana-to-Kanji conversion. The recognition subsystem uses a hybrid algorithm consisting of a patt ern matcher and a neural network discriminator. Special care was taken to improve the recognition of non-Kanji and simple Kanji characters f requently used in fast data entry. The post-processor predicts consecu tive characters on the basis of bigrams modified by the addition of pa rts of speech and substitution of macro characters for Kanji character s. A Kana-to-Kanji conversion method designed for ease of use with a p en interface has also been integrated into the system. In an experimen t in which 2,900 samples of Kanji and non-Kanji characters were obtain ed from 20 subjects, it was observed that the original recognition acc uracy of 83.7% (the result obtained by using the pattern matching reco gnizer) was improved to 90.7% by adding the neural network discriminat or, and that it was further improved to 94.4% by adding the post-proce ssor. The improved recognition accuracy for non-Kanji characters was p articularly marked.