Tweet
Title: Estimation of joint torques and contact forces from video in rock climbing
postdoctoral position 2019-2020
Laboratories: LJK/GIPSA-Lab
INRIA/CNRS/Grenoble-INP/UGA FRANCE
Objectives:
This postdoctoral position will explore new approaches in rock climbing motion analysis in order to
estimate joint torques and contact forces from video. It follows previous works by Quaine et al. 2017 on
an optimization method for simultaneous prediction of contact forces and joint torques from optical data.
In Biomechanics, the usual experimental procedure to estimate muscle torques at joints requires to
measure both kinematic data of the body from motion capture and contact forces with the environment
from force sensors. While accurate, this approach is complex to deploy and limits the ability to address
dynamics analysis in an environment such as rock-climbing mainly because of the absence of contact
force sensors at each hold. Markerless methods are now expanding rapidly to allow an easier approach
for kinematic analysis of body motion (OpenPose, Cao et al., 2017). In this project, we are interested in
investigating further optimization procedure coupled with a markerless analysis in order to include
estimation of joint torques and contact forces in the specific case of rock climbing.
Context:
The context of the project gathers the expertise of the Math-Sports group in Applied Mathematics and
Computer Sciences at the LJK laboratory and the Biomechanics team at GIPSA laboratory, both in
Grenoble area, France. The project will benefit from infrastructure offered at the laboratories, such as
high-end motion capture system and 3D dynamic scanning. In addition, a close collaboration with the
French National Team in Climbing (FFME) will allow to investigate the application of the new methods
in the context of training of high-level athletes.
Position application:
This position will be at the LJK laboratory, Grenoble, France. The LJK is a joint laboratory between
INRIA, CNRS, Grenoble-INP and UGA. Candidate should hold a PhD related to Biomechanics,
Computer Graphics or Computer Vision. Programming skills in C++ or Python are required. Knowledge
of 3D modeling software such as Autodesk Maya or Blender is appreciated. Applications are to be sent
to both Lionel Reveret and Franck Quaine (see contact below). The duration is for one year, with a
possible extension of 6 months. Net salary is 2500 euros per month, starting in 2019.
Contact:
Lionel Reveret, LJK, lionel.reveret@inria.fr
Franck Quaine, GIPSA-Lab, franck.quaine@gipsa-lab.grenoble-inp.fr
Bibliography:
Cao, Z., Simon, T., Wei, S.E., Sheikh, Y., Realtime multi-person 2d pose estimation using part affinity
fields, CVPR 2017.
Quaine, F., L. Reveret, L., Courtemanche, S., Kry, P., "Postural regulation and motion simulation in
rock climbing". The science of climbing and mountaineering, Routledge, pp.111-128, 2017, Routledge
Research in Sport and exercise Science, 2017, www.motionworkshop.org/climbing_forces.mp4
postdoctoral position 2019-2020
Laboratories: LJK/GIPSA-Lab
INRIA/CNRS/Grenoble-INP/UGA FRANCE
Objectives:
This postdoctoral position will explore new approaches in rock climbing motion analysis in order to
estimate joint torques and contact forces from video. It follows previous works by Quaine et al. 2017 on
an optimization method for simultaneous prediction of contact forces and joint torques from optical data.
In Biomechanics, the usual experimental procedure to estimate muscle torques at joints requires to
measure both kinematic data of the body from motion capture and contact forces with the environment
from force sensors. While accurate, this approach is complex to deploy and limits the ability to address
dynamics analysis in an environment such as rock-climbing mainly because of the absence of contact
force sensors at each hold. Markerless methods are now expanding rapidly to allow an easier approach
for kinematic analysis of body motion (OpenPose, Cao et al., 2017). In this project, we are interested in
investigating further optimization procedure coupled with a markerless analysis in order to include
estimation of joint torques and contact forces in the specific case of rock climbing.
Context:
The context of the project gathers the expertise of the Math-Sports group in Applied Mathematics and
Computer Sciences at the LJK laboratory and the Biomechanics team at GIPSA laboratory, both in
Grenoble area, France. The project will benefit from infrastructure offered at the laboratories, such as
high-end motion capture system and 3D dynamic scanning. In addition, a close collaboration with the
French National Team in Climbing (FFME) will allow to investigate the application of the new methods
in the context of training of high-level athletes.
Position application:
This position will be at the LJK laboratory, Grenoble, France. The LJK is a joint laboratory between
INRIA, CNRS, Grenoble-INP and UGA. Candidate should hold a PhD related to Biomechanics,
Computer Graphics or Computer Vision. Programming skills in C++ or Python are required. Knowledge
of 3D modeling software such as Autodesk Maya or Blender is appreciated. Applications are to be sent
to both Lionel Reveret and Franck Quaine (see contact below). The duration is for one year, with a
possible extension of 6 months. Net salary is 2500 euros per month, starting in 2019.
Contact:
Lionel Reveret, LJK, lionel.reveret@inria.fr
Franck Quaine, GIPSA-Lab, franck.quaine@gipsa-lab.grenoble-inp.fr
Bibliography:
Cao, Z., Simon, T., Wei, S.E., Sheikh, Y., Realtime multi-person 2d pose estimation using part affinity
fields, CVPR 2017.
Quaine, F., L. Reveret, L., Courtemanche, S., Kry, P., "Postural regulation and motion simulation in
rock climbing". The science of climbing and mountaineering, Routledge, pp.111-128, 2017, Routledge
Research in Sport and exercise Science, 2017, www.motionworkshop.org/climbing_forces.mp4