Stereoscopic tracking during a zooming period is facilitated to alleviate eye fatigue. An image processing device includes an imaging unit 14 that acquires a stereoscopic image including a plurality of viewpoint images, an operation unit 16 that acquires a zoom value, an electronic zoom processing unit 17 that performs magnification of the stereoscopic image by image processing on the basis of the acquired zoom value, a monitor 21 capable of outputting the magnified stereoscopic image, and a control unit 25 that outputs the stereoscopic image immediately before or immediately after a change in the zoom value to the monitor 21 as a stereoscopic still image magnified by the electronic zoom processing unit 17 while the acquired zoom value is varying, and outputs the stereoscopic image acquired by the imaging unit 14 to the monitor 21 as a stereoscopic moving image while the zoom value is not varying.