An Approach for Human Machine Interaction using Dynamic Hand Gesture Recognition

Gesture recognition is one of the most challenging area of research in computer vision. The interface for human computer interaction can easily be created using the gesture recognition. This paper suggests a novel as well as robust approach of dynamic hand gesture recognition for the human machine interaction system. Here, the system is designed to work in real time with images captured through web camera during experimentation. In RGB and HSV color space, skin color modeling is done for segmenting geometrical approximation of hand using contours. The region of interest (ROI) is used to locate the hand within the image. Fingers are counted using the contour defects and centroid tracking is used to track the hand over the sequence of frames. In experiments, the eight types of dynamic hand gestures are performed by user and these gestures are successfully recognized by the system. This proposed methodology is functioning admirably in term of recognition accuracy up to 95 % for dynamic hand gesture recognition.