Real-time human pose recognition in parts from single depth images

Shotton, Jamie; Fitzgibbon, Andrew; Cook, Mat; Sharp, Toby; Finocchio, Mark; Moore, Richard B.; Kipman, Alex; Blake, A.

doi:10.1109/cvpr.2011.5995316

Cited by 3,215 publications

(2,368 citation statements)

References 29 publications

Supporting

Mentioning

2,322

Contrasting

Unclassified

Order By: Relevance

“…Also during our experiments we observed that a curtain, the operating bed, a server rack and a stand for an ultrasound device were infrequently detected as a person. To overcome these shortcomings, in future iterations the Microsoft Kinect V2 camera with the Microsoft algorithms [26] will be used. The results show that moving persons in the operating theater can be detected in very short time; however, it has to be taken into account that the data are adjusted using the delay (rounded 1 s) resulting in higher values for detection time.…”

Section: Discussionmentioning

confidence: 99%

Time-of-flight-assisted Kinect camera-based people detection for intuitive human robot cooperation in the surgical operating room

et al. 2015

View full text Add to dashboard Cite

Background Scene supervision is a major tool to make medical robots safer and more intuitive. The paper shows an approach to efficiently use 3D cameras within the surgical operating room to enable for safe human robot interaction and action perception. Additionally the presented approach aims to make 3D camera-based scene supervision more reliable and accurate. Methods A camera system composed of multiple Kinect and time-of-flight cameras has been designed, implemented and calibrated. Calibration and object detection as well as people tracking methods have been designed and evaluated. Results The camera system shows a good registration accuracy of 0.05 m. The tracking of humans is reliable and accurate and has been evaluated in an experimental setup using operating clothing. The robot detection shows an error of around 0.04 m. Conclusions The robustness and accuracy of the approach allow for an integration into modern operating room. The data output can be used directly for situation and workflow detection as well as collision avoidance.

show abstract

Section: Discussionmentioning

confidence: 99%

Time-of-flight-assisted Kinect camera-based people detection for intuitive human robot cooperation in the surgical operating room

et al. 2015

View full text Add to dashboard Cite

show abstract

“…We use all frames of actor 1 to construct the test set (≈ 62k poses). The H36M skeleton includes some spurious joints that we delete, which results in the same 20 joints present in the Kinect skeleton [20]. All frames are given in relative xyz coordinates centered at the hip node, unless otherwise stated.…”

Section: Methodsmentioning

confidence: 99%

“…Reasoning about human pose is a key ingredient in recent successful applications of computer vision systems [20]. Accurately capturing the variability of human pose is challenging because there is both a variation between different persons as well as a combinatorial number of possible poses a single person can assume.…”

Section: Introductionmentioning

confidence: 99%

A Non-parametric Bayesian Network Prior of Human Pose

Lehrmann

Gehler

Nowozin

2013

2013 IEEE International Conference on Computer Vision

View full text Add to dashboard Cite

Having a sensible prior of human pose is a vital ingredient for many computer vision applications, including tracking and pose estimation. While the application of global non-parametric approaches and parametric models has led to some success, finding the right balance in terms of flexibility and tractability, as well as estimating model parameters from data has turned out to be challenging. In this work, we introduce a sparse Bayesian network model of human pose that is non-parametric with respect to the estimation of both its graph structure and its local distributions. We describe an efficient sampling scheme for our model and show its tractability for the computation of exact log-likelihoods. We empirically validate our approach on the Human 3.6M dataset and demonstrate superior performance to global models and parametric networks. We further illustrate our model's ability to represent and compose poses not present in the training set (compositionality) and describe a speed-accuracy trade-off that allows realtime scoring of poses.

show abstract

“…Another reason which inconsistent positions are estimated because it's not tracked. Artificial vibration is result of the acquisition error from camera which produces unwanted vibrations on the extracted joint-position and makes length of bones change during the motion [8]. They apply object recognition approach that produce designing an intermediate body parts representation that maps the difficult pose estimation problem into a simpler per-pixel classification problem.…”

Section: Analysis Of "Le Sas"mentioning

confidence: 99%

“…Then generate confidence-scored 3D proposals of several body joints by reprojecting the classification result and finding local modes. They use consumer hardware which runs at 200 frames per second, this allow to show high accuracy on both synthetic and real test sets, and investigates the effect of several training parameters [8]. Self-occlusion, (Top-right) Bone-length variation [7].…”

Section: Analysis Of "Le Sas"mentioning

confidence: 99%

Dynamic Sensor Selection Based on Joint Data Quality in the Context of a Multi-Kinect Module inside the CAVE “Le SAS”

Salous¹,

Newton

Leroy

et al. 2016

IJCTE

View full text Add to dashboard Cite

Abstract-In order to create interaction, immersive VR rooms can receive various types of inputs from a user. One of these input types are the gestures and movements of the user in the CAVE, which can be captured with sensors such as the Microsoft Kinect. However, the large scale of an immersive room implies the use of multiple Kinects to provide optimal coverage. This multi-Kinect set-up requires an effective selection method to determine the most accurate Kinects depending on the user's position. The example of VR environment described in this paper is a CAVE named "Le SAS" that currently supports 4 Kinects. This paper describes the SAS's features as well as the constraints the implemented solution had to take into account. It will justify the technical choices and provide experiment results.

show abstract

Real-time human pose recognition in parts from single depth images

Cited by 3,215 publications

References 29 publications

Time-of-flight-assisted Kinect camera-based people detection for intuitive human robot cooperation in the surgical operating room

Time-of-flight-assisted Kinect camera-based people detection for intuitive human robot cooperation in the surgical operating room

A Non-parametric Bayesian Network Prior of Human Pose

Dynamic Sensor Selection Based on Joint Data Quality in the Context of a Multi-Kinect Module inside the CAVE “Le SAS”

Contact Info

Product

Resources

About