The inclusion of more potentially correct words in the candidate sets is important to improve the accuracy of Large Vocabulary Continuous Speech Recognition （LVCSR）. A candidate expansion algorithm based on the Weig...The inclusion of more potentially correct words in the candidate sets is important to improve the accuracy of Large Vocabulary Continuous Speech Recognition （LVCSR）. A candidate expansion algorithm based on the Weighted Syllable Confusion Matrix （WSCM） is proposed. First, WSCM is derived from a confusion network. Then, the reeognised candidates in the confusion network is used to conjeeture the most likely correct words based on WSCM, after which, the conjectured words are combined with the recognised candidates to produce an expanded candidate set. Finally, a combined model having mutual information and a trigram language model is used to rerank the candidates. The experiments on Mandarin film data show that an improvement of 9.57% in the character correction rate is obtained over the initial recognition performance on those light erroneous utterances.展开更多
基金supported by the National Natural Science Foundation of China under Grants No.61005004,No.61175011,No.61171193the Next-Generation Broadband Wireless Mobile Communications Network Technology Key Project under Grant No.2011ZX03002-005-01+2 种基金the One Church,One Family,One Purpose(111Project)under Grant No.B08004the Key Project of Ministry of Science and Technology of China under Grant No.2012ZX-03002019-002the National High Techni-cal Research and Development Program of China(863Program)under Grant No.2011A-A01A205
文摘The inclusion of more potentially correct words in the candidate sets is important to improve the accuracy of Large Vocabulary Continuous Speech Recognition （LVCSR）. A candidate expansion algorithm based on the Weighted Syllable Confusion Matrix （WSCM） is proposed. First, WSCM is derived from a confusion network. Then, the reeognised candidates in the confusion network is used to conjeeture the most likely correct words based on WSCM, after which, the conjectured words are combined with the recognised candidates to produce an expanded candidate set. Finally, a combined model having mutual information and a trigram language model is used to rerank the candidates. The experiments on Mandarin film data show that an improvement of 9.57% in the character correction rate is obtained over the initial recognition performance on those light erroneous utterances.
