Research Of Gesture Recognition Based On Finger Angle Characteristics

Nowadays, the computer technology has been widely used, it not only helpspeople reduce the heavy manual labor, but also saves a lot of labor time because of the highprocessing rate, promotes the development of new technologies. People’s life has be graduallyimproved, and they pursuit a more comfortable life, human-computer interactiontechnology is adapted to this demand, has become an important part of people’s life. Gesturerecognition based on vision is one of the key technology of human-computer interaction, isbecoming a hot spot of research in this field. However, because different people have made ina different time and place of the gestures will be very different, and has different meaning indifferent cultural background, the vision based gesture recognition has becomea multidisciplinary topic，so that the machine instead of the human brain to make the rightresponse to the gesture is also challenging.Study on the two aspects this paper mainly from the skin color segmentation and gesturerecognition method. Because of the background image in the practical application of thecomplex, this paper proposes an algorithm of new skin color segmentation. In the module isrealized by YCbCr and Lab two color space modeling the skin color segmentation, therebyincreasing the skin and non-skin color region difference and overcome the YCbCr space oversensitive to red, yellow two color. Gesture recognition, this paper proposes a new algorithm ofgesture recognition. Connected the detected finger and the center of palm, forming a modelsimplified hand gesture, the relative angle and the number of finger tips are classified on thebasis of the gestures for gesture analysis and recognition. The experimental results showthat: the algorithm is high accuracy and has good real-time performance, which is notinfluenced by the direction and scale of gesture.