Research On Moving Hand Tracking And Hand Track Recognition Technologies

In recent years, with the development of computer application technology, the human-computer interaction based on computer vision technology has been widely applied gradually. As an important part of human computer interaction, dynamic tracking and recognition of human hands has got widespread attention at home and abroad. The human-computer interaction technology based on gestures takes hand and hand’s trajectory as object recognition, providing a good human-computer interaction interface for communication between smart devices, such as computers.The dynamic identification study can be divided into three steps:hand detection segmentation, tracking and dynamic hand trajectory’s recognition. Hand detection procedure test and accurately segment the hand area of video according to the certain characteristics of a man’s hand; Dynamic hand tracking extracts center of hand in continuous image sequences and gets the movement trajectory of a man’s hand; Finally, pattern recognition technology can identify the semantics of human’s hand movement,By setting gestures simulating commands of mouse trajectory, a new method is proposed based on the combination of background difference and color histogram, improving the efficiency and accuracy of the human detection; Against the existing shortcomings of various gestures trajectory tracking methods, proposed method combining the characteristics with hand movement direction, achieved fast gesture trajectory extraction. The effectiveness of the proposed method is verified by simulation and the superiority; Based on TDPCA and high-performance extreme learning machine, a new gesture trajectory identification algorithm is put forward. Firstly main feature is extracted from samples of trajectory using TDPCA, and then, the input weights and bias value of in hidden layer neurons should be randomly assigned. Finally, by calculating and outputting weights, EELM classifying model after training is used to classify and realize hand trajectory semantic identification. The simulations verified that the new method has higher recognition rate on the premise of guarantee real-time requirements.