Design And Implementation Of Dynamic Gesture Recognition Based On Multi-motion Sensors

With the continuous development of the people-oriented concept in the computer field,the development of human-computer interaction technology is an important development direction of computer application research.Among them,more and more gesture interactions are an important way of human-computer interaction.Through research on sensing devices,sensing methods,and recognition methods,gesture interaction technology has been continuously developed.In this paper,motion sensor is used as the acquisition device of dynamic gesture raw data.Because the detection accuracy of single motion sensor is not high,and there are problems such as noise interference and calculation error accumulation,the result obtained by gesture recognition algorithm often has a large error with the actual gesture.In order to recognize gestures more effectively,this paper synchronizes the motion information of gestures by four motion sensors,acquires gesture motion trajectories,normalization and down-sampled value processing,which simplifies the recognition process and reduces the recognition difficulty.The hidden Markov model is fused into multiple observation sequences for establishing a gesture model.In the process of dynamic gesture recognition,the gesture is first divided into several atomic gestures,and the pre-processed motion information is used to generate the feature quantity of the atomic gesture.A discrete hidden Markov model is then built based on these feature quantities.The Baum-Welch algorithm is mainly used to train the model,and the interpolation method is used to solve the problem of insufficient training data.Finally,the trained model is used to identify atomic gestures,and these atomic gestures are combined into complex,coherent gestures.Through the experimental verification of multi-motion sensor gesture recognition method and system,multi-sensor based gesture recognition is effective for complex and consistent gesture recognition,and the system cost is relatively low.As a human-computer interaction mode,it can be integrated with other systems and has high applicability and extensibility.