PROBLEM TO BE SOLVED: To provide a method, a device, and a program for multilingual speech recognition and a method, a device, and a program for multilingual speaker adaptation that secure high speech recognition precision and operate at a high speed even when a non-native speaker speaks words of one language.;SOLUTION: There are provided the method, device, and program for multilingual speech recognition which are equipped with pronunciation dictionaries 102 of a plurality of languages to be recognized and corresponding sound models 103, one multilingual grammar 104 for recognition generated by using words defined in the pronunciation dictionaries 102 of the plurality of languages to be recognized, and one speech analysis part 101 and a search part 105 independent of the languages; and the search part 105 uses the sound models of the plurality of languages to be recognized for feature parameters extracted by the speech analysis part 101 from an input speech and collates them with a word string represented by the grammar for recognition to output the work string having the largest collation score as a recognition result.;COPYRIGHT: (C)2004,JPO
展开▼
