Non-trajectory-based gesture recognition in human-computer interaction based on hand skeleton data

Currently, no efficient, accurate and flexible gesture recognition algorithm has been developed to recognize non-trajectory-based gesture recognition. Therefore, we aim to construct a gesture recognition algorithm to not only complete gesture recognition accurately and quickly but also adapt to individual differences. In this paper, we present a novel non-trajectory-based gesture recognition method (NT-GRM) based on hand skeleton information and a hidden Markov model (HMM). To recognize a static gesture, the direction information of each bone section of the hands was taken as the observation data to construct the HMM. In addition, multiple static gestures were detected in turn to identify a dynamic gesture. As determined by experimental verification, the NT-GRM can complete recognition in a system containing ten interactive gestures with a recognition accuracy of over 95% and a recognition speed of 21.73 ms. The training time required for each static gesture model is 2.56 s. And the NT-GRM can identify static and dynamic gestures accurately and quickly with small training samples in different functional modes. In conclusion, the NT-GRM can be applied to the development of gesture interaction systems to help developers realize practical functions such as gesture library construction, user gesture customization, and user gesture adaptation.