Gaussian hand gesture recognition based mobilitydevice controller

The development and investigation of alternative mobility device control is presented in this work. The system uses 2D visual information, which is acquired from an ordinary web-cam, and controls the electrical drives of the mobility device by tracking and recognizing the gestures of the hand. Hand tracking is achieved by using an algorithm, which combines two methods: a statistical Gaussian method and a discrete Fourier transformation. Proposed algorithm is adaptive and flexible allowing utilizing unique gesture commands which depend on person's motor abilities. Experimental investigation proves the stable robustness, performance and high accuracy of the proposed mobility controller.