APPARATUS, METHOD, AND PROGRAM FOR TITLE SIGNAL PROCESSING

摘要

PROBLEM TO BE SOLVED: To properly detect a title a deviating in presentation timing to a voice.

SOLUTION: A voice recognition part 141 performs voice recognition of utterance contents from a voice signal and detects the utterance timing of the utterance contents. A subtitle analysis part 142 recognizes title contents from a title signal and detects the presentation timing of the title contents. A coincidence retrieval part 143 detects coincidence between the utterance contents and title contents. Here, when the coincidence is detected, a deviation quantity decision part 144 calculates the time difference between the detected utterance timing and presentation timing of the coincident utterance contents and title contents.

COPYRIGHT: (C)2005,JPO&NCIPI