| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a technology which finds the importance level of an object from an image pick-up pattern and classifies images for each object. <P>SOLUTION: At an object importance level setting part 11, an object importance level weight parameter is preliminarily set to determine the importance level of the object as a pick-up object. At an object group table generation part 12, at least one image group is inputted from a user to extract an object for each image, and the extracted object is correlated to the extraction origin image information of each object and image pick-up pattern information as an importance level indicator, to generate an object data group table. At an important object extraction part 13, the importance level of each object is determined based on the object data group and the importance level weight parameter thus set, and an important object is extracted. At an object kind classification part 14, the important objects of the same kind thus extracted are classified as the same objects, and a tentative name for each kind is set. At an object classification part 15, the object data group is classified for each kind of the object thus set according to the classification of the important object. <P>COPYRIGHT: (C)2005,JPO&NCIPI |
