An image processing apparatus that enables to prolong lives of device parts of which lives are limited by preventing unnecessary returns from the power saving mode, and reduce unnecessary power consumption. A first living body detection unit detects a living body that approaches the apparatus to input a trigger. A second living body detection unit detects a living body that approaches the apparatus for other purposes. A switching unit switches a mode from a first electric power mode to a second electric power mode when the first living body detection unit detects the living body and when the second living body detection unit does not detect the living body, and maintains the first electric power mode when the first living body detection unit detects the living body and when the second living body detection unit detects the living body.