A system for accurately detecting a patient's movement during imaging procedures includes a camera for providing a stream of camera images of a part of a patient's exterior and a fiducial element mountable on part of the patient's exterior. The fiducial element is detectable in the stream of images. An image processor is configured to detect a displacement of the fiducial element based on consecutive images including in at least the stream of images and to generate an output signal indicative of the displacement. The fiducial element has an in-plane stiffness which is substantially larger than an in-plane stiffness of the part of the patient's exterior. In addition, the fiducial element and the part of the patient's exterior are provided with substantially equal outer in-plane dimensions.