The New Type Of Human-computer Interaction System Based On Wearable Device

Independent of the mouse, keyboard and other existing traditional human-computer interaction devices, more flexible and natural human-computer interaction has always been a hot spot. In recent years, emerging intelligent and convenient wearable devices has led the future development of science and technology. The human-computer interaction system based on wearable devices will have broad application prospects.Hand gesture interaction is an important branch of human-computer interaction, according to different gesture modeling methods, the human-computer interaction based on hand gestures includes based on mechanical and electronic sensor signals or based on computer vision and so on. The way based on mechanical and electronic sensor signals mainly has the problem such as equipment direct contact with the hand, constraint hand movement. While traditional gesture interaction based on computer vision because of the camera in the face of the user, exist complicated background, and other parts of the body color interference, be sheltered, limit problem such as the distance between camera and people. In view of the limitations of traditional gesture recognition, this article uses wearable camera and acceleration sensor as a data acquisition device, use camera to collect images, and acceleration sensor to remove jitter, it don’t need user’s hand direct contact with the equipment, also can be a very good solution to solve the above problems of the traditional visual gesture interaction.This paper study gesture interaction system based on wearable device, including the system of the illumination brightness change processing, head shaking interfere processing, the hardware architecture of parallel algorithm implementation and high speed data cache handling, etc. The main work and results were as follows:1) On the algorithm combine the lighting processing algorithm with the improved adaptive method of skin color segmentation model to solve the problem that the system illumination brightness change interfere hands segmentation extraction, put forward using acceleration sensor to get the head acceleration and tilt angle change value to solve the problem that the jitter caused by the head cause the system output gestures coordinates jitter;2) On the system hardware design uses the FPGA(Field-Programmable Gate Array) parallel compute camera image data, uses ARM(Advanced RISC Machine) processor as system USB output interface, research optimize system hardware architecture of parallel implementation of algorithm, high-speed data cache processing. 3) Complete roaming and click of the gesture mouse, test the recognition rate and interaction experiment, can finish the 60 frames per second for 800 * 600 resolution of video image high-speed high-performance real-time gesture recognition and tracking...