Accuracy and robustness of Kinect pose estimation in the context of coaching of elderly population

Obdržálek, Štěpán; Kurillo, Gregorij; Ofli, Ferda; Bajcsy, Růžena; Seto, Edmund; Jimison, Holly; Pavel, M.

doi:10.1109/embc.2012.6346149

Cited by 274 publications

(156 citation statements)

References 6 publications

Supporting

Mentioning

148

Contrasting

Unclassified

Order By: Relevance

“…Another main drawback of the Kinect skeleton is in a very non-anthropometric kinematic model with variable limb lengths. Although this does not seem to be a problem for ergonomic assessment by OWAS and others applications where less information is needed (Obdrzálek et al, 2012), the use of other methods such as RULA, REBA, LULA or PATH (Buchholz et al, 1996) requires to measure joint rotations and an improved kinematic model. MicrosoftÔ recently announced the launch of Kinect 2Ô.…”

Section: Discussionmentioning

confidence: 99%

Using Kinect™ sensor in observational methods for assessing postures at work

2014

View full text Add to dashboard Cite

a b s t r a c tThis paper examines the potential use of KinectÔ range sensor in observational methods for assessing postural loads. Range sensors can detect the position of the joints at high sampling rates without attaching sensors or markers directly to the subject under study. First, a computerized OWAS ergonomic assessment system was implemented to permit the data acquisition from KinectÔ and data processing in order to identify the risk level of each recorded postures. Output data were compared with the results provided by human observers, and were used to determine the influence of the sensor view angle relative to the worker. The tests show high inter-method agreement in the classification of risk categories (Proportion agreement index ¼ 0.89 k ¼ 0.83) when the tracked subject is facing the sensor. The camera's point of view relative to the position of the tracked subject significantly affects the correct classification of the postures. Although the results are promising, some aspects involved in the use of low-cost range sensors should be further studied for their use in real environments.

show abstract

Section: Discussionmentioning

confidence: 99%

Using Kinect™ sensor in observational methods for assessing postures at work

2014

View full text Add to dashboard Cite

show abstract

“…The accuracy and robustness of Kinect based rehabilitation technique have been tested on [7], where they have presented the problems which occur due to noise and occlusion while capturing data with Kinect. In order to overcome this problem we have proposed to use a custom made posture database (PDB) with which we can replace the missing information from the data that we have captured.…”

Section: Related Workmentioning

confidence: 99%

Distance Physical Rehabilitation System Framework with Multi-Kinect Motion Captured Data

Rafiuzzaman¹,

Öz²

2015

CAE

View full text Add to dashboard Cite

“…Since the algorithm for marker recognition is still under development, the Kinect-based orientation measurement accuracy could not be evaluated on the field. Therefore, it was deduced from information and data available in the literature [35], [36].…”

Section: B Sensor Characterizationmentioning

confidence: 99%

Design and performance analysis of an indoor position tracking technique for smart rollators

Nazemzadeh

Fontanelli

Macii

et al. 2013

International Conference on Indoor Positioning and Indoor Navigation

View full text Add to dashboard Cite

Abstract-This paper presents a position tracking technique based on multisensor data fusion for rollators helping elderly people to move safely in large indoor spaces such as public buildings, shopping malls or airports. The proposed technique has been developed within the FP7 project DALi, and relies on an extended Kalman filter processing data from dead-reckoning sensors (i.e. encoders and gyroscopes), a short-range radio frequency identification (RFID) system and a front Kinect camera. As known, position tracking based on dead-reckoning sensors only is intrinsically affected by growing uncertainty. In order to keep such uncertainty within wanted boundaries, the position values are occasionally updated using a coarse-grained grid of low-cost passive RFID tags with known coordinates in a given map-based reference frame. Unfortunately, RFID tag detection does not provide any information about the orientation of the rollator. Therefore, a front camera detecting some markers on the walls is used to adjust direction. Of course, the data rate from both the RFID reader and the camera is not constant, as it depends on the actual user's trajectory and on the distance between pairs of RFID tags and pairs of markers. Therefore, the average distance between tags and markers should be properly set to achieve a good trade-off between overall deployment costs and accuracy. In the paper, the results of a simulation-based performance analysis are reported in view of implementing the proposed localization and tracking technique in a real environment.

show abstract

Accuracy and robustness of Kinect pose estimation in the context of coaching of elderly population

Cited by 274 publications

References 6 publications

Using Kinect™ sensor in observational methods for assessing postures at work

Using Kinect™ sensor in observational methods for assessing postures at work

Distance Physical Rehabilitation System Framework with Multi-Kinect Motion Captured Data

Design and performance analysis of an indoor position tracking technique for smart rollators

Contact Info

Product

Resources

About