Research And Application Of Virtual And Reality Interaction Method Based On Biometrics

Biometric technology refers to a technology that uses a computer's inherent physiological characteristics(fingerprint,iris,face,blood vessels,etc.)to perform personal identification.Augmented reality is a comprehensive technology that uses computer-generated additional information to enhance or expand the real scenes observed by users.It is a research hotspot in recent years.Organically combining the two technologies is a lot of industrial research Popular.The two technologies of biometrics and augmented reality are effectively fused and applied to device terminals,and gesture recognition based on computer vision is used as a medium to realize the interaction between virtuality and reality.Therefore,this paper aims at the real-time functional requirements of the fusion system of face recognition and augmented reality technology in biometrics,using Kinect sensors as computer vision input devices,a virtual and real interactive system based on biometrics is proposed,and the identity is determined by face recognition Real-time interaction using dynamic gesture recognition and augmented reality technology.Through the test and analysis of face recognition algorithms based on geometric features,sparse representation,BP neural network,and convolutional neural network,it is known that the method based on geometric features has poor recognition effect on blocked face images The sparse representation-based method is prone to redundancy in the dictionary learning phase.The BP neural network-based method has a slower convergence rate,and it is prone to solve the contradiction between the application case size and the network scale.The sharing of convolution kernel parameters in the hidden layers and the sparseness of the connections between layers enable the convolutional neural network to maintain a stable effect with a small amount of calculation.Comprehensively comparing the advantages and disadvantages of various algorithms,the convolutional neural network algorithm is selected as the user identification algorithm based on the biometric virtual and real interactive system in this paper.The DTW algorithm is selected for dynamic gesture recognition.In view of the problems existing in the traditional DTW algorithm,a combination of redundant template modeling and optimization of correlation distortion calculation is proposed to improve the DTW algorithm.In the template matching and calculation stage,the traditional DTW algorithm uses the exhaustive method to perform template matching,which requires a large amount of calculation and has defects in the response time of real-time gesture recognition.Therefore,the redundant template modeling method is introduced to change the gesture template search strategy,thereby increasing the accuracy of gesture matching and making the human-computer interaction have a smoother experience.The virtual reality interaction system based on biometrics is designed and implemented.It is proposed to integrate face recognition,gesture recognition and augmented reality interaction,and apply the improved DTW algorithm to the system.Therefore,the real-time and accuracy of gesture recognition are improved,and the augmented reality interactive system based on biometric recognition is suitable for indoor and outdoor scenes.