TY - JOUR
T1 - Parsing human skeletons in an operating room
AU - Belagiannis, Vasileios
AU - Wang, Xinchao
AU - Shitrit, Horesh Beny Ben
AU - Hashimoto, Kiyoshi
AU - Stauder, Ralf
AU - Aoki, Yoshimitsu
AU - Kranzfelder, Michael
AU - Schneider, Armin
AU - Fua, Pascal
AU - Ilic, Slobodan
AU - Feussner, Hubertus
AU - Navab, Nassir
N1 - Funding Information:
This work was supported in part by the Swiss National Science Foundation and by DFG - Deutsche Forschungsgemeinschaft under the project “Advanced Learning for Tracking and Detection in Medical Workflow Analysis”. The authors would like to thank Iro Laina for helping with the data preparation.
Publisher Copyright:
© 2016, Springer-Verlag Berlin Heidelberg.
PY - 2016/10/1
Y1 - 2016/10/1
N2 - Multiple human pose estimation is an important yet challenging problem. In an operating room (OR) environment, the 3D body poses of surgeons and medical staff can provide important clues for surgical workflow analysis. For that purpose, we propose an algorithm for localizing and recovering body poses of multiple human in an OR environment under a multi-camera setup. Our model builds on 3D Pictorial Structures and 2D body part localization across all camera views, using convolutional neural networks (ConvNets). To evaluate our algorithm, we introduce a dataset captured in a real OR environment. Our dataset is unique, challenging and publicly available with annotated ground truths. Our proposed algorithm yields to promising pose estimation results on this dataset.
AB - Multiple human pose estimation is an important yet challenging problem. In an operating room (OR) environment, the 3D body poses of surgeons and medical staff can provide important clues for surgical workflow analysis. For that purpose, we propose an algorithm for localizing and recovering body poses of multiple human in an OR environment under a multi-camera setup. Our model builds on 3D Pictorial Structures and 2D body part localization across all camera views, using convolutional neural networks (ConvNets). To evaluate our algorithm, we introduce a dataset captured in a real OR environment. Our dataset is unique, challenging and publicly available with annotated ground truths. Our proposed algorithm yields to promising pose estimation results on this dataset.
KW - Human pose estimation
KW - Medical workflow analysis
KW - Part-based model
UR - http://www.scopus.com/inward/record.url?scp=84979287781&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84979287781&partnerID=8YFLogxK
U2 - 10.1007/s00138-016-0792-4
DO - 10.1007/s00138-016-0792-4
M3 - Article
AN - SCOPUS:84979287781
SN - 0932-8092
VL - 27
SP - 1035
EP - 1046
JO - Machine Vision and Applications
JF - Machine Vision and Applications
IS - 7
ER -