出版社：The Institute of Image Information and Television Engineers
摘要：A problem of the conventional monitoring system for closed captions was that no alarm was given to operators if the broadcast closed captions did not correspond with the program audio. We propose the use of an automatic monitoring system of closed captions that lets the operator know if the captions and audio no longer correspond. This system compares the closed captions on the air with the audio text data obtained using speech recognition technology to find whether the two data correspond. We found that this system performed well in forty-one programs.
