Hand gesture based interface for aiding visually impaired

Hand gesture recognition is a growing and very vast field of research. Numerous work have been done and a lot of work still remains to be done for providing a intuitive, innovative and natural way of non verbal communication, which is more familiar to human beings. Gesture Recognition is widely used in sign language, alternative computer interfaces, Immersive game technology etc. The aim of this paper is to present a system for hand gesture recognition to provide a interface for aiding visually impaired users on the basis of detection of some useful shape based features like orientation, area, centroid, extrema, location, presence of fingers and thumb in image. The approach discussed in this paper solely depends on the shape of the hand gesture. It does not comprise color or texture of the image, which are variant to different light shades and other influences. This approach uses some pre-processing steps for removal of background noise and employs K-means clustering for segmentation of hand object so that only segmented hand objects or cluster is to be processed in terms of shape based features. This unique approach can recognize around 36 different gestures on the bases of 7 bit binary sequence or string generated as a output of this algorithm. The proposed implemented approach has been tested on 360 images, and it gives approximate recognition rate of 94%. One of the great benefits of this algorithm is that it takes only fractional part of a second to recognize the hand gesture which makes it computationally efficient as compare to the other existing approach. The proposed algorithm is simple and independent of user characteristics. And also it does not require any kind of training of data like in HMM or neural network.