Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

KinePose: A temporally optimized inverse kinematics technique for 6DOF human pose estimation with biomechanical constraints

Gildea, Kevin LU ; Mercadal-Baudart, Clara ; Blythman, Richard ; Smolic, Aljosa and Simms, Ciaran (2022) 24th Irish Machine Vision and Image Processing Conference p.105-112
Abstract
Computer vision/deep learning-based 3D human pose estimation methods aim to localize human joints from images and videos. Pose representation is normally limited to 3D joint positional/translational degrees of freedom (3DOFs), however, a further three rotational DOFs (6DOFs) are required for many potential biomechanical applications. Positional DOFs are insufficient to analytically solve for joint rotational DOFs in a 3D human skeletal model. Therefore, we propose a temporal inverse kinematics (IK) optimization technique to infer joint orientations throughout a biomechanically informed, and subject-specific kinematic chain. For this, we prescribe link directions from a position-based 3D pose estimate. Sequential least squares quadratic... (More)
Computer vision/deep learning-based 3D human pose estimation methods aim to localize human joints from images and videos. Pose representation is normally limited to 3D joint positional/translational degrees of freedom (3DOFs), however, a further three rotational DOFs (6DOFs) are required for many potential biomechanical applications. Positional DOFs are insufficient to analytically solve for joint rotational DOFs in a 3D human skeletal model. Therefore, we propose a temporal inverse kinematics (IK) optimization technique to infer joint orientations throughout a biomechanically informed, and subject-specific kinematic chain. For this, we prescribe link directions from a position-based 3D pose estimate. Sequential least squares quadratic programming is used to solve a minimization problem that involves both frame-based pose terms, and a temporal term. The solution space is constrained using joint DOFs, and ranges of motion (ROMs). We generate 3D pose motion sequences to assess the IK approach both for general accuracy, and accuracy in boundary cases. Our temporal algorithm achieves 6DOF pose estimates with low Mean Per Joint Angular Separation (MPJAS) errors (3.7°/joint overall, & 1.6°/joint for lower limbs). With frame-by-frame IK we obtain low errors in the case of bent elbows and knees, however, motion sequences with phases of extended/straight limbs results in ambiguity in twist angle. With temporal IK, we reduce ambiguity for these poses, resulting in lower average errors. Code and supplementary material are available. (Less)
Please use this url to cite or link to this publication:
author
; ; ; and
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
Human pose estimation, Computer vision, Inverse kinematics, Motion capture, Biomechanics
host publication
Proceedings of the 24th Irish Machine Vision and Image Processing conference (IMVIP 2022)
pages
105 - 112
publisher
Irish Pattern Recognition & Classification Society
conference name
24th Irish Machine Vision and Image Processing Conference
conference location
Belfast, United Kingdom
conference dates
2022-08-31 - 2022-09-02
ISBN
978-0-9934207-7-1
language
English
LU publication?
no
id
5dd5a4fe-a5f5-4a57-aa52-a5ca5884ed6b
alternative location
https://arxiv.org/abs/2207.12841
date added to LUP
2022-12-13 20:14:18
date last changed
2023-06-09 11:21:55
@inproceedings{5dd5a4fe-a5f5-4a57-aa52-a5ca5884ed6b,
  abstract     = {{Computer vision/deep learning-based 3D human pose estimation methods aim to localize human joints from images and videos. Pose representation is normally limited to 3D joint positional/translational degrees of freedom (3DOFs), however, a further three rotational DOFs (6DOFs) are required for many potential biomechanical applications. Positional DOFs are insufficient to analytically solve for joint rotational DOFs in a 3D human skeletal model. Therefore, we propose a temporal inverse kinematics (IK) optimization technique to infer joint orientations throughout a biomechanically informed, and subject-specific kinematic chain. For this, we prescribe link directions from a position-based 3D pose estimate. Sequential least squares quadratic programming is used to solve a minimization problem that involves both frame-based pose terms, and a temporal term. The solution space is constrained using joint DOFs, and ranges of motion (ROMs). We generate 3D pose motion sequences to assess the IK approach both for general accuracy, and accuracy in boundary cases. Our temporal algorithm achieves 6DOF pose estimates with low Mean Per Joint Angular Separation (MPJAS) errors (3.7°/joint overall, & 1.6°/joint for lower limbs). With frame-by-frame IK we obtain low errors in the case of bent elbows and knees, however, motion sequences with phases of extended/straight limbs results in ambiguity in twist angle. With temporal IK, we reduce ambiguity for these poses, resulting in lower average errors. Code and supplementary material are available.}},
  author       = {{Gildea, Kevin and Mercadal-Baudart, Clara and Blythman, Richard and Smolic, Aljosa and Simms, Ciaran}},
  booktitle    = {{Proceedings of the 24th Irish Machine Vision and Image Processing conference (IMVIP 2022)}},
  isbn         = {{978-0-9934207-7-1}},
  keywords     = {{Human pose estimation; Computer vision; Inverse kinematics; Motion capture; Biomechanics}},
  language     = {{eng}},
  pages        = {{105--112}},
  publisher    = {{Irish Pattern Recognition & Classification Society}},
  title        = {{KinePose: A temporally optimized inverse kinematics technique for 6DOF human pose estimation with biomechanical constraints}},
  url          = {{https://arxiv.org/abs/2207.12841}},
  year         = {{2022}},
}