Motion Estimation Using Physical Simulation

Damien Duff; Jeremy Wyatt; Rustam Stolkin

doi:10.1109/ROBOT.2010.5509590

Motion Estimation Using Physical Simulation

Damien Duff, Jeremy Wyatt, Rustam Stolkin

Computer Science

Research output: Contribution to conference (unpublished) › Paper › peer-review

10 Citations (Scopus)

210 Downloads (Pure)

Abstract

We consider the task of monocular visual motion estimation from video image sequences. We hypothesise that performance on the task can be improved by incorporating an understanding of physically likely and feasible object dynamics. We test this hypothesis by incorporating a physical simulator into a least-squares estimation procedure. We initialise a full trajectory estimate using RANSAC followed by gradient descent refinement. We present results for 2D image sequences consisting of single ambiguous, visible or occluded balls, as well as results for 3D computer-generated sequences of objects in free-flight with added noise. Results suggest that restricting the estimation to allow only motions that are feasible according to the physics simulator can produce marked improvement when the observed object motion is within the limits of the physics simulator and its world model. Conversely, merely penalising deviations from feasible physical dynamics produces a consistent but incremental improvement over more common dynamics models.

Original language	English
Pages	1511-1517
Number of pages	7
DOIs	https://doi.org/10.1109/ROBOT.2010.5509590
Publication status	Published - 7 May 2010
Event	Proceedings of the 2010 IEEE International Conference on Robotics and Automation (ICRA 2010) - Duration: 7 May 2010 → …

Conference

Conference	Proceedings of the 2010 IEEE International Conference on Robotics and Automation (ICRA 2010)
Period	7/05/10 → …

Access to Document

10.1109/ROBOT.2010.5509590

BouncingBalls_ICRA2010_VF11Preprint.pdf

http://dx.doi.org/10.1109/ROBOT.2010.5509590

Cite this

@conference{bdad9054110a437596362a97cc5e5ba3,

title = "Motion Estimation Using Physical Simulation",

abstract = "We consider the task of monocular visual motion estimation from video image sequences. We hypothesise that performance on the task can be improved by incorporating an understanding of physically likely and feasible object dynamics. We test this hypothesis by incorporating a physical simulator into a least-squares estimation procedure. We initialise a full trajectory estimate using RANSAC followed by gradient descent refinement. We present results for 2D image sequences consisting of single ambiguous, visible or occluded balls, as well as results for 3D computer-generated sequences of objects in free-flight with added noise. Results suggest that restricting the estimation to allow only motions that are feasible according to the physics simulator can produce marked improvement when the observed object motion is within the limits of the physics simulator and its world model. Conversely, merely penalising deviations from feasible physical dynamics produces a consistent but incremental improvement over more common dynamics models.",

author = "Damien Duff and Jeremy Wyatt and Rustam Stolkin",

year = "2010",

month = may,

day = "7",

doi = "10.1109/ROBOT.2010.5509590",

language = "English",

pages = "1511--1517",

note = "Proceedings of the 2010 IEEE International Conference on Robotics and Automation (ICRA 2010) ; Conference date: 07-05-2010",

}

TY - CONF

T1 - Motion Estimation Using Physical Simulation

AU - Duff, Damien

AU - Wyatt, Jeremy

AU - Stolkin, Rustam

PY - 2010/5/7

Y1 - 2010/5/7

N2 - We consider the task of monocular visual motion estimation from video image sequences. We hypothesise that performance on the task can be improved by incorporating an understanding of physically likely and feasible object dynamics. We test this hypothesis by incorporating a physical simulator into a least-squares estimation procedure. We initialise a full trajectory estimate using RANSAC followed by gradient descent refinement. We present results for 2D image sequences consisting of single ambiguous, visible or occluded balls, as well as results for 3D computer-generated sequences of objects in free-flight with added noise. Results suggest that restricting the estimation to allow only motions that are feasible according to the physics simulator can produce marked improvement when the observed object motion is within the limits of the physics simulator and its world model. Conversely, merely penalising deviations from feasible physical dynamics produces a consistent but incremental improvement over more common dynamics models.

AB - We consider the task of monocular visual motion estimation from video image sequences. We hypothesise that performance on the task can be improved by incorporating an understanding of physically likely and feasible object dynamics. We test this hypothesis by incorporating a physical simulator into a least-squares estimation procedure. We initialise a full trajectory estimate using RANSAC followed by gradient descent refinement. We present results for 2D image sequences consisting of single ambiguous, visible or occluded balls, as well as results for 3D computer-generated sequences of objects in free-flight with added noise. Results suggest that restricting the estimation to allow only motions that are feasible according to the physics simulator can produce marked improvement when the observed object motion is within the limits of the physics simulator and its world model. Conversely, merely penalising deviations from feasible physical dynamics produces a consistent but incremental improvement over more common dynamics models.

U2 - 10.1109/ROBOT.2010.5509590

DO - 10.1109/ROBOT.2010.5509590

M3 - Paper

SP - 1511

EP - 1517

T2 - Proceedings of the 2010 IEEE International Conference on Robotics and Automation (ICRA 2010)

Y2 - 7 May 2010

ER -

Motion Estimation Using Physical Simulation

Abstract

Conference

Access to Document

Fingerprint

Cite this