Detection of patient movement for video EEG monitoring

We propose a method to monitor automatically the motion of an epileptic during video EEG monitoring. The method consists of two procedures. First, it segments a head image into objects which are head, mouth and eyes. Second, the spatial local optimization method is employed for the estimation of optical flow, which quantifies the motion. Then, we obtain the averaged optical flow and set a threshold level to detect the motion event of the objects, respectively. Finally, when the averaged optical flow is more than the threshold level, the motion events of each objects are decided. In the results, we show that the optical flow represents the detailed motion of a patient