| 摘要 |
PROBLEM TO BE SOLVED: To automatically expand a specific image among content images. SOLUTION: The content response processing method is composed of a storage means 110 for storing a speech feature amount contained in the speech to learn and a code book 110CB in which an enhanced state probability is made to correspond to a calm state probability; a speech feature amount extracting means 120 for extracting the speech feature amount contained in speech data; an enhanced state likelihood calculating means 130 for calculating the enhanced stated likelihood using the enhanced state probability extracted from the code book, based on the extracted speech feature amount; an utterance state determining means 150 for judging whether an analyzed interval has been in the enhanced state, based on the enhanced state likelihood; and an image magnification means 160 for magnifying a content image in the interval for determining by a predetermined magnification to output the image, when the interval has been in the enhanced state. COPYRIGHT: (C)2006,JPO&NCIPI
|
