| 摘要 |
<P>PROBLEM TO BE SOLVED: To generate a speaking image expressing natural mouth movements for any given word. <P>SOLUTION: An image generator 123 arrays sets of basic mouth shape image data each representing a different basic mouth shape in the order of appearance, when pronouncing an object word input by a user, of an initial mouth shape that is formed in advance of pronouncing a prescribed sound, is different from the mouth shape matching the vowel of that sound and is one of a plurality of basic mouth shapes and of a final mouth shape that is a mouth shape formed when the pronunciation of one sound ends and is one of a plurality of basic mouth shapes and, by interpolating intermediate image data between two consecutive sets of basic mouth shape image data by using these two sets of basic mouth shape image data, generates a speaking image expressing movements of the mouth pronouncing the object word. This principle is applicable to, for instance, a device for the training of lip reading. <P>COPYRIGHT: (C)2012,JPO&INPIT |
