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 indi...

Full description

Saved in:
Bibliographic Details
Published in:Multimedia tools and applications 2022-06, Vol.81 (15), p.20509-20539
Main Authors: Jia, Lesong, Zhou, Xiaozhou, Xue, Chengqi
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Bones
Computer Communication Networks
Computer Science
Data Structures and Information Theory
Electromyography
Gesture recognition
Human-computer interaction
Interactive systems
Markov chains
Multimedia
Multimedia Information Systems
Semantics
Special Purpose and Application-Based Systems
Training
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary: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.
ISSN:1380-7501
1573-7721
DOI:10.1007/s11042-022-12355-8