When composition of a first captured image having a tilted subject image is corrected, the first image is rotated such that the subject image stands in an erect position, according to the attitude of the camera achieved during capture of the first image. Images corresponding to missing portions, which arise in the image area for recording due to rotation, are extracted from a second captured image as compensation images. The second image is obtained by capturing an image of the same subject as that captured by the first image, at the different angle of view and attitude of the camera. An image obtained by enlarging the compensation images extracted from the second image according to the angle of view employed for capturing the first image is synthesized with the rotated first image, to thus make the subject image stand in an erect position and to obtain a corrected image without involving a decreased number of pixels.