For each of a plurality of second images other than a first image in an image group having a plurality of images, a feature point pair is generated by associating a second feature point of the second image with a first feature point of the first image based on a feature amount of the second feature point. A feature point pair is detected from the generated feature point pairs, where a position of the second feature point in the detected feature point pair is located within a predetermined region. A region including first feature points of the first image is extracted, where in the extracted region, detection counts of the feature point pairs exceed a predetermined threshold.