.png){kind=icon}
.png){kind=icon}
Return
Multi-View Real-Time Human Motion Recognition Based on Ensemble Learning
.png){kind=icon}
PRE
AI
DOI：10.1109/JSEN.2021.3094548.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper deals with the real-time recognition from multiple spatial angles of concealed human motions with ultra wide band (UWB) through-the-wall radar (TWR). To conquer the performance loss incurred by diverse human motion in a single view, a multi-view real-time human motion recognition model based on ensemble learning is proposed. Specifically, we first proposes a multi-view human motion recognition model based on Stacking parallel ensemble learning algorithm, which is used to realize the real-time human motion recognition based on UWB TWR. Secondly, in order to address the irrationality of the existing accuracy evaluation criteria to evaluate the real-time motion recognition algorithm, a real-time model evaluation criterion based on normalized harmonic weighted intersection over union (NHW-IOU) is proposed. Finally, the collected multi-view human motion data are used to verify the effectiveness of the proposed algorithm. The actual measurement results show that the average recognition performance of the proposed Stacking model has improved by 14.76% compared with the single-view model, which is of great significance for using multi-view data to improve network performance. Moreover, compared with the bi-directional long short-term memory (Bi-LSTM) and Gated Recurrent Unit (GRU) models, the proposed model has better performance in accuracy and time delay.
Keywords:
Real-time systems
Radar
Target recognition
Stacking
Ultra wideband radar
Time-frequency analysis
Data models
Human motion recognition
ultra wide band through-the-wall radar
multi-view
stacking model
ensemble learning
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.5
Papers:
2.1W
Citations:
7.3W
Organization
No organization information available