Vision-based sleep mode detection for a smart TV

Sleep mode detection is one of the significant features of power management and green computing. However, for a television or a smart TV, it is difficult to detect a deactivation event because the user can use these devices without input from an input device. We propose a robust method to detect deactivation events based on a vision approach involving face detection and motion detection for a smart TV. Experiments are performed on a large dataset. The proposed approach significantly reduces false detections of faces and complement missed humans by means of motion detection.