Hasso-Plattner-InstitutSDG am HPI
Hasso-Plattner-InstitutDSG am HPI
Login
 

Human Motion Analysis Using 3D Cameras

Chair for Digital Health - Connected Healthcare
Hasso Plattner Institute

Office: Campus III Building G2, Room G-2.1.20
Tel.: +49 331 5509-4853
Email: Justin.albert(at)hpi.de
Links: Homepage

Supervisor: Prof. Dr. Bert Arnrich

Starting Date: 01.10.2019

 

In my research, I focus on human motion analysis using primarily 3D cameras but also other sensor modalities such as Inertial Measurement Units and Electrocardiography. The projects range from using a low-cost 3D camera for gait analysis to predicting subjective exertion in strength training. In the following sections, I want to give an overview of the past projects of the last year and my current research. 

Evaluation of the Pose Tracking Accuracy of the Microsoft Kinect v2 and Azure Kinect Camera

Microsoft released the first version of its Kinect camera in 2010 as a gaming controller for the Xbox gaming console. It can track certain joint positions of users in 3D. It combines an RGB camera with a 3D depth sensor. Since the second camera generation, the Time-of-Flight (ToF) principle has been used for depth estimation. This method estimates the depth by emitting IR-light into the scene and measuring the time until it gets reflected and returns to the sensor. For 3D motion tracking, Kinect v2 used randomized decision forests to estimate the joint locations, as described in [4]. In 2019, a new Kinect generation, Azure Kinect, was released where the focus is shifting away from games towards industrial applications. The skeleton tracking algorithm utilizes deep learning with Convolutional Neural Networks (CNN) to estimate the human poses. The research community has used the Kinect camera for medical and biomedical applications and analysis for many years. 

In this project, we utilized the latest Microsoft Azure Kinect camera for gait analysis. Gait analysis is an essential tool for the early detection of neurological diseases and assessing the risk of falling in elderly people. More specifically, we evaluated the pose tracking performance of the Azure Kinect camera compared to its predecessor Kinect v2 in treadmill walking. We have used a Vicon multi-camera motion capturing system and the 39 marker Plug-in Gait model as the gold standard. Five young and healthy subjects walked on a treadmill at three different velocities. Data were recorded simultaneously with all three camera systems. To compare the spatial agreement of joint locations, we have developed an external camera calibration to spatially align the 3D skeleton data from both Kinect cameras and the Vicon system. Specific gait parameters were calculated for all three camera systems, including step length, step time, step width, and stride time. The results showed that the improved hardware and the motion tracking algorithm of the Azure Kinect camera led to significantly higher accuracy of the spatial gait parameters than the predecessor Kinect v2. At the same time, no significant differences were found between the temporal parameters. The results of this study were published in the MDPI sensors journal. 

References

  1. Gunnar Borg. “Perceived exertion as an indicator of somatic stress.” In: Scandinavian journal of rehabilitation medicine (1970).

  2. E. Barsoum, J. Kender, and Z. Liu, “HP-GAN: Probabilistic 3D Human Motion Prediction via GAN,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2018.

  3. M. Capecci, M. G. Ceravolo, F. Ferracuti, S. Iarlori, A. Monteri`u, L. Romeo, and F. Verdini, “The kimore dataset: Kinematic assessment of movement and clinical scores for remote monitoring of physical rehabilitation,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 27, no. 7, pp. 1436–1448, July 2019.
  4. Shotton, J.; Fitzgibbon, A.; Cook, M.; Sharp, T.; Finocchio, M.; Moore, R.; Kipman, A.; Blake, A. Real-time Human Pose Recognition in Parts from Single Depth Images 2011. pp. 1297–1304. doi:10.1109/CVPR.2011.5995316.