Keyword recognition and extraction by multiple-LVCSRs with 60, 000 words in speech-driven WEB retrieval task.
Masahiko MatsushitaHiromitsu NishizakiSeiichi NakagawaTakehito UtsuroPublished in: INTERSPEECH (2004)
Keyphrases
- keywords
- speech corpus
- continuous speech recognition
- text recognition
- spoken words
- recognition rate
- recognition engine
- word recognition
- website
- object recognition
- speech recognition systems
- web documents
- automatic speech recognition
- automatic transcription
- recognition accuracy
- pattern recognition
- n gram
- web pages
- objects in cluttered scenes
- information extraction
- web applications
- action recognition
- noisy environments
- information access
- speech recognition
- document analysis
- multi lingual
- keyword extraction
- data extraction
- spontaneous speech
- wordnet
- semantic web
- recognition errors
- numeral strings
- search engine
- speech sounds
- word sense disambiguation
- web mining