To accurately and quickly detect and follow-up a specific image from frame images to be successively obtained.

Until the first specific image (face image) is detected, a template is scanned according to a fixed path. After the face image is detected, a priority area corresponding to the detection area is set in a frame image. In the next frame, the detection processing of the face image is performed only in the set priority area, and the detection processing of the face image is operated about the residual area. The face image detection processing when the priority area is set is provided with a fixed upper limit time. After the start of the detection processing of the face image, the face image detection processing is interrupted with respect to the frame image when the upper limit time is reached. As for the next frame image, after the face image detection processing is performed in the priority area, the face image detection is resumed from the previously interrupted position and following position.