HMM/ANN System for Vietnamese Continuous Digit Recognition(Artificial Intelligence III)

スポンサーリンク

概要

• 論文の詳細を見る

• The study of a system for Vietnamese continuous digit recognition is described. The CSLU Toolkit was used to develop and implement hybrid HMM/ANN recognition systems. Experiments were done with a corpus of 442 sentences with 2340 words, which were extracted from two telephone-speech corpora: "22 Language vl.2" and "Multi-Language Telephone Speech vl.2". In our experiments, a context-dependent phoneme recognizer has achieved better recognition performance than a context-dependent demi-syllable recognizer and a context-independent phoneme recognizer. Among feature sets applied to the context-dependent phoneme recognizer, the set of 12 PLP features with CMS, energy and corresponding delta values has achieved the best recognition result (96.83% word accuracy and 87.67% sentence correct).

• 一般社団法人情報処理学会の論文
• 2004-12-04

著者

• Duc Dang

  Alcatel Nsv Vietnam Post And Telecommunication

• Thang Vu

  Institute Of Information Technology Vietnamese Academy Of Science And Technology

• Hosom John-paul

  Center For Spoken Language Understanding (cslu-ogi)

• MAI LUONG

  Institute of Information Technology, Vietnamese Academy of Science and Technology

• Mai Luong

  Institute Of Information Technology Vietnamese Academy Of Science And Technology

• Dang Ngoc

  Alcatel NSV, Vietnam Post and Telecommunication

• Luong Chi

  Institute of Information Technology, Vietnamese Academy of Science and Technology

関連論文

• HMM/ANN System for Vietnamese Continuous Digit Recognition(Artificial Intelligence III)
• HMM/ANN System for Vietnamese Continuous Digit Recognition(Additional Presentation)(Joint Workshop of Vietnamese Society of AI, SIGKBS-JSAI, ICS-IPSJ, and IEICE-SIGAI on Active Mining)

スポンサーリンク