J.L. Raheja, R. Shyam, and U. Kumar (India)
Hand Gesture, Human-machine interaction, Motion detection, Object recognition and Image segmentation, Image matching.
This paper presents a fast and efficient technique for recognition of continuous hand gestures with a stationary background from video stream. The system consists of four modules. A real time hand tracking and extraction, feature extraction, Principle component analysis (PCA) algorithm training, and gesture recognition. First of all, real-time hand tracking and extraction algorithm is applied to trace the moving hand and to extract the hand region, and then motion analysis is performed to characterize the temporal features. Instead of capturing each frame and tracking hand movement for gesture recognition, it captures only that frame for analysis which shows least motion of the hand. That is, the frame is captured where hand stops moving for a few moments of time. This instance of the frame is where the gesture of the hand is complete. In this way the proposed technique becomes fast and memory efficient. In this proposed technique we capture 5 frames per second from video stream. After that we compare 3 continuous frames to know the frame, containing static posture shown by hand. This static posture is recognized as a hand gesture. Now it is matched with stored gesture database to know its meaning. Before matching gesture, preprocessing is performed to remove other part of human body and to select only part of hand showing some useful gesture.
Important Links:
Go Back
