S control: Accelerometer-based gesture recognition for media control

Wearable devices such as Samsung Galaxy Gear enable intuitive interaction with media devices via hand gestures. The hand gestures are detected using the inertial sensors present in wearable devices. Existing sensor based gesture detection algorithms have limitations which prevent its use in smart wearable devices and smart phones. There are two major limitations, one is gesture spotting in a stream of continuous sensor data and another is the amount of training required to arrive at acceptable gesture detection accuracy. We discuss these and other limitations in detail and propose an algorithm to overcome them. Further the paper discusses a method of finding the intensity of the performed gesture which can be used to improve the user experience.