中文漢語語音辨識系統

This invention is to provide a kind of speech recognition technology for the Chinese languages, including several dialects, e.g. Mandarin, Taiwanese (Min-nan), and Hakka. By using the data structure of the Chinese character tree, multiple pronunciations could be arranged in the same node of a particular Chinese character, which is usually uttered in different pronunciations by people from different dialectical areas, such that the technological difficulties of pronunciation varieties for the Chinese languages could be solved in a unified manner. The point of this invention is the use of the Chinese character tree, which adopts the Chinese characters instead of syllables as the node to build a lexicon tree, where all the possible pronunciations of different dialects can be considered at the same time.