An image shooting device with an image sensing portion that acquires a series of input images formed in chronological order; a face detection portion that, detects a human face from each input image outputting position information of the detected face; a priority face choice portion that chooses a priority face from the detected faces; an identification portion that, searches for the priority face; and a re-choice portion that, when the priority face is detected from a first input image acquired but the priority face is not detected by the identification portion from a second input image retains information including position information executes a re-choice processing to re-choose the priority face searching within the third input image for a face to be re-chosen as a priority face.