In an image processing apparatus, the head top of a subject is detected, and the positions of a camera and a mirror are adjusted so that the coordinates of the detected head top coincide with the coordinates of a reference head top. When it is determined that the coordinates coincide with each other, the image size of a face of the subject is detected, and the zoom of the camera is controlled so that the detected image size and a reference image size coincide with each other. When it is determined that the image size of the face and the reference image size coincide, the positions of the camera and the mirror are adjusted again so that the coordinates of the head top of the subject and the coordinates of the reference head top coincide with each other.