A speech recognition system of continuously spoken Japanese sentences and an application to a speech input device

A speech recognition system for continuously spoken Japanese sentences is described. The system has a knowledge source, in which syntactic, semantic and vocabulary knowledge are contained. It can treat not only simple sentences but also simple complex and compound sentences without any restriction regarding the tasks. Phoneme identification is performed in real time, and after a slight revision has been carried out the speech sound is represented by a pseudo-phoneme string. In the syntactic knowledge system, inflections of inflectional words are taken into account. The syntactic analyzer, using a syntactic state transition network, predicts the following words and outputs their syntactic interpretations. The semantic knowledge system deals with the meanings of each word, in particular, it deals with case structures for verbs. The semantic knowledge is used for the prediction of the following words and construction of a semantic network of the input story. Moreover, a slightly modified version can be used as a speech input device of a word processor.