作者: Fang Lu , Janani Janakiraman , Susan M. Cox
DOI:
关键词: Closed captioning 、 Speech recognition 、 Audio signal 、 Stress (linguistics) 、 Speaker diarisation 、 Pronunciation 、 Videoconferencing 、 Computer science
摘要: A method for automatically enabling closed captioning in video conferencing when a heavy accent is detected from current speaker provided. Language background and/or ethnicity information received as user preference. An acceptable level determined according to the audio signal of speaking language received. pronunciation compared with standard language. determined, and level. If comparison determines that does not comply level, enabled signal. complies