摘要 |
<P>PROBLEM TO BE SOLVED: To simply distinguish categories for subjects in images, such as whose face or facial expression, when a subject is a face. <P>SOLUTION: An image included with a face is inputted (S201), and a plurality of local features are detected from the inputted image. Using the plurality of detected local features, a facial area in the image is specified (S202). Based on the deviation of the detection result of each local feature in the facial area and a detection result pre-acquired as standard for each local feature in the face area, facial expression is distinguished (S204). <P>COPYRIGHT: (C)2005,JPO&NCIPI |