There is provided an image processing device including an image acquisition part acquiring an image; a depth acquisition part acquiring a depth in association with a pixel included in the image; a target object detection part detecting a region of a predetermined target object in the image; a target object detection distance selection part selecting the depth corresponding to the pixel included in the detected region as a target object detection distance; a local maximum distance selection part selecting the depth having a local maximum frequency in a frequency distribution of the depths as a local maximum distance; and a determination part determining whether the image is a target object image obtained by shooting the target object depending on whether a degree of closeness between a value of the target object detection distance and a value of the local maximum distance is higher than a predetermined value.