Wiktionary as a Source for Automatic Pronunciation Extraction
作者： Sebastian Ochs , Tim Schlippe , Tanja Schultz
DOI:
关键词: Natural language processing 、 Speech synthesis 、 International Phonetic Alphabet 、 Word (computer architecture) 、 Pronunciation 、 German 、 Computer science 、 Word error rate 、 Artificial intelligence
摘要: In this paper, we analyze whether dictionaries from the World Wide Web which contain phonetic notations, may support rapid creation of pronunciation within speech recognition and synthesis system building process. As a representative dictionary, selected Wiktionary [1] since it is at hand in multiple languages and, addition to definitions words, many notations terms International Phonetic Alphabet (IPA) are available. Given word lists four English, French, German, Spanish, calculated percentage words with Wiktionary. Furthermore, two quality checks were performed: First, compared pronunciations based on GlobalPhone project, had been created rule-based fashion manually cross-checked [2]. Second, analyzed impact automatic (ASR) systems. French achieved best coverage, containing 92.58% for list as well 76.12% 30.16% country international city names. our ASR systems evaluation, Spanish gained most improvement 7.22% relative error rate reduction.
isca-speech.org参考文章(13)
Marelie H. Davel, Etienne Barnard, The efficient generation of pronunciation dictionaries: human factors during bootstrapping. conference of the international speech communication association. ,(2004)
Tanja Schultz, GlobalPhone: A Multilingual Speech and Text Database developed at Karlsruhe University international conference on spoken language processing. ,(2002)
Matthias Wölfel, Channel selection by class separability measures for automatic transcriptions on distant microphones. conference of the international speech communication association. pp. 582- 585 ,(2007)
Xiaojin Zhu, R. Rosenfeld, Improving trigram language modeling with the World Wide Web international conference on acoustics, speech, and signal processing. ,vol. 1, pp. 533- 536 ,(2001) , 10.1109/ICASSP.2001.940885
Kevin A. Lenzo, Vincent Pagel, Alan W. Black, Issues in Building General Letter to Sound Rules SSW. pp. 77- 80 ,(1998)
Ariadna Font Llitjós, Alan W. Black, Evaluation and collection of proper name pronunciations online language resources and evaluation. ,(2002)
T. Schultz, A. Waibel, Polyphone decision tree specialization for language adaptation international conference on acoustics, speech, and signal processing. ,vol. 3, pp. 1707- 1710 ,(2000) , 10.1109/ICASSP.2000.862080
John Kominek, Alan W Black, Learning Pronunciation Dictionaries: Language Complexity and Word Selection Strategies language and technology conference. pp. 232- 239 ,(2006) , 10.3115/1220835.1220865
Arnab Ghoshal, Martin Jansche, Sanjeev Khudanpur, Michael Riley, Morgan Ulinski, WEB-derived pronunciations international conference on acoustics, speech, and signal processing. pp. 4289- 4292 ,(2009) , 10.1109/ICASSP.2009.4960577
Tanja Schultz, Alex Waibel, Language-independent and language-adaptive acoustic modeling for speech recognition Speech Communication. ,vol. 35, pp. 31- 51 ,(2001) , 10.1016/S0167-6393(00)00094-7