Static Hand Gesture Recognition Using Haar-Like Features

Hand gesture recognition plays a crucial role in communication between human and computer or robot. It is used to improve Human-Computer Interaction (HCI) for the sake of making the communication more natural and much easier. Static hand gesture or posture recognition using Haar-like features is being presented in this paper. Two static hand gestures which are index finger and fist are trained using Haar-like features algorithm. Index finger represents left click mouse event while fist represents right click mouse event. AdaBoost algorithm is applied in the training phase to increase accuracy and robustness of the system. Since this is a real-time system, built-in webcam is used to capture the image of the gesture. Brightness and distance are tested for evaluation of this system. Some static imported images are also tested. The experimental results show that both static hand gestures achieve the highest accuracy under a high degree (80%-100%) of brightness. Index finger and fist achieve 90.4% and 91.2% accuracy respectively under a high degree of brightness. The best distance is 80cm from the screen. Index finger achieves 92% accuracy for 80cm distance while the fist achieves 95.2% for both 80cm and 100cm distances