中国邮电高校学报(英文) ›› 2021, Vol. 28 ›› Issue (3): 86-94.doi: 10.19682/j.cnki.1005-8885.2021.0011

• Artificial Intelligence • 上一篇    下一篇

Real-time hand tracking based on YOLOv4 model and Kalman filter

杜绪伟,陈东,刘华江,马兆昆,杨倩倩   

  1. 青岛科技大学
  • 收稿日期:2020-07-05 修回日期:2020-12-03 出版日期:2021-06-30 发布日期:2021-06-22
  • 通讯作者: 陈东 E-mail:duxuqust@163.com

Real-time hand tracking based on YOLOv4 model and Kalman filter

  • Received:2020-07-05 Revised:2020-12-03 Online:2021-06-30 Published:2021-06-22

摘要:

Aiming at the shortcomings of current gesture tracking methods in accuracy and speed, based on deep learning You Only Look Once version 4 (YOLOv4) model, a new YOLOv4 model combined with Kalman filter rea-time hand tracking method was proposed. The new algorithm can address some problems existing in hand tracking technology such as detection speed, accuracy and stability. The convolutional neural network (CNN) model YOLOv4 is used to detect the target of current frame tracking and Kalman filter is applied to predict the next position and bounding box size of the target according to its current position. The detected target is tracked by comparing the estimated result with the detected target in the next frame and, finally, the real-time hand movement track is displayed. The experimental results validate the proposed algorithm with the overall success rate of 99.43%

at speed of 41.822 frame/ s, achieving superior results than other algorithms. 


关键词:

hand tracking, You Only Look Once version 4 (YOLOv4) model, Kalman filter, real-time

Abstract:

Aiming at the shortcomings of current gesture tracking methods in accuracy and speed, based on deep learning You Only Look Once version 4 (YOLOv4) model, a new YOLOv4 model combined with Kalman filter rea-time hand tracking method was proposed. The new algorithm can address some problems existing in hand tracking technology such as detection speed, accuracy and stability. The convolutional neural network (CNN) model YOLOv4 is used to detect the target of current frame tracking and Kalman filter is applied to predict the next position and bounding box size of the target according to its current position. The detected target is tracked by comparing the estimated result with the detected target in the next frame and, finally, the real-time hand movement track is displayed. The experimental results validate the proposed algorithm with the overall success rate of 99.43%

at speed of 41.822 frame/ s, achieving superior results than other algorithms. 

Key words:

hand tracking, You Only Look Once version 4 (YOLOv4) model, Kalman filter, real-time