| 摘要 |
According to the present invention, the amount of computation required for image recognition processing can be reduced by extracting only image recognition learning information for an object that may appear in a region having the geographical property of a current position and comparing the image recognition learning information with ambient-image information. |
