Robust single finger movement detection scheme for real time wheelchair control by physically challenged people

Abstract

With the proliferation of human-computer interfaces that rely on user gestures, there exists a real opportunity to develop systems to ease the lives of disabled as well as normal people. Driving the wheelchair by hand may not be feasible for a physically challenged person with limited hand movement and as an alternate, simple finger movement based control system can assist them a great deal. In this paper, a robust scheme of finger movement detection is proposed for the direction control of a wheelchair. In order to obtain the fingertip location, a finger movement detection scheme is developed based on background subtraction and some morphological operations, where computationally expensive classifiers with trainers are avoided. The proposed method is capable of identifying finger movement in real time from images captured even by a very low resolution off-the-shelf cell phone camera. The control operation between the cell-phone and wheelchair is carried out wirelessly with the help of bluetooth technology. It is found that the proposed method is suitable under varying light intensity and noise conditions. Extensive analyses of various effects, such as shade, camera position, and speed of finger movement have been performed and a very high level of accuracy with low computational cost in terms of real-time precise direction control is achieved. The proposed approach could have significant impact by providing means to the physically challenged section in the society to become self-reliant as well as contribute for the overall development and turn themselves into a strength of the society.