| 摘要 |
<P>PROBLEM TO BE SOLVED: To improve recognition accuracy when detecting an object or a person from an image. <P>SOLUTION: An area in which the object is photographed is extracted from an image to be processed, and a feature quantity of the image in the area is calculated. Matching processing is performed using the calculated feature quantity, a feature quantity (parameter) prepared beforehand for the object to be recognized, and probability values (context parameters) among objects possibly photographed simultaneously in one image and among objects possibly photographed over a plurality of images. The matching result is provided to a user as a recognition result. The present invention is applicable to a recognition unit for recognizing an object. <P>COPYRIGHT: (C)2007,JPO&INPIT |
