A Dynamic Gesture Recognition Method Based On Computer Vision

The development of science and technology is changing with each passing day,and the way of human-computer interaction has undergone several changes.A wide range of graphical interfaces and touch interaction are not the most suitable ways for human interaction.Therefore,human-computer interaction based on computer vision is gradually shifting from the academic research field to the modern application field.In this thesis,a dynamic gesture recognition method based on computer vision is proposed,and a music player based on dynamic gesture recognition is realized by this method.The dynamic gesture recognition method based on computer vision mainly includes dynamic gesture detection and segmentation,dynamic gesture tracking and dynamic gesture recognition.A hybrid Gaussian skin color model based on brightness index combined with frame difference method is proposed for dynamic gesture detection.The experiment shows that this method has a higher success rate than the method,which use the threshold segmentation method based on HSV color space combined with frame difference method.A camshift algorithm based on kalman filter is proposed for dynamic gesture tracking.The experiment shows that this method overcomes the shortcomings of camshift algorithm,which is sensitive to light changes and vulnerable to large area skin color.This thesis uses the hidden Markov model to train and recognize the dynamic gesture command.The experiment shows that this method has achieved the expected result in the recognition of dynamic gesture command.Finally,a music player based on dynamic gesture recognition is implemented in this thesis.The test results shows that the dynamic gesture recognition method proposed in this thesis has obtained good experimental results.