An adaptive-scale robust estimator for motion estimation

Trung Ngo Thanh, Hajime Nagahara, Ryusuke Sagawa, Yasuhiro Mukaigawa, Masahiko Yachida, Yasushi Yagi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Although RANSAC is the most widely used robust estimator in computer vision, it has certain limitations making it ineffective in some situations, such as the motion estimation problem, in which uncertainty on the image features changes according to the capturing conditions. The greatest problem is that the threshold used by RANSAC to detect inliers cannot be changed adaptively; instead it is fixed by the user. An adaptive scale algorithm must therefore be applied in such cases. In this paper, we propose a new adaptive scale robust estimator that adaptively finds the best solution with the best scale to fit the inliers, without the need for predefined information. Our new adaptive scale estimator matches the residual probability density from an estimate and the standard Gaussian probability density function to find the best inlier scale. Our algorithm is evaluated in several motion estimation experiments under varying conditions and the results are compared with several of the latest adaptive-scale robust estimators.

Original languageEnglish
Title of host publication2009 IEEE International Conference on Robotics and Automation, ICRA '09
Pages2455-2460
Number of pages6
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event2009 IEEE International Conference on Robotics and Automation, ICRA '09 - Kobe, Japan
Duration: May 12 2009May 17 2009

Other

Other2009 IEEE International Conference on Robotics and Automation, ICRA '09
CountryJapan
CityKobe
Period5/12/095/17/09

Fingerprint

Motion estimation
Probability density function
Computer vision
Experiments
Uncertainty

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Software
  • Artificial Intelligence
  • Electrical and Electronic Engineering

Cite this

Thanh, T. N., Nagahara, H., Sagawa, R., Mukaigawa, Y., Yachida, M., & Yagi, Y. (2009). An adaptive-scale robust estimator for motion estimation. In 2009 IEEE International Conference on Robotics and Automation, ICRA '09 (pp. 2455-2460). [5152445] https://doi.org/10.1109/ROBOT.2009.5152445

An adaptive-scale robust estimator for motion estimation. / Thanh, Trung Ngo; Nagahara, Hajime; Sagawa, Ryusuke; Mukaigawa, Yasuhiro; Yachida, Masahiko; Yagi, Yasushi.

2009 IEEE International Conference on Robotics and Automation, ICRA '09. 2009. p. 2455-2460 5152445.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Thanh, TN, Nagahara, H, Sagawa, R, Mukaigawa, Y, Yachida, M & Yagi, Y 2009, An adaptive-scale robust estimator for motion estimation. in 2009 IEEE International Conference on Robotics and Automation, ICRA '09., 5152445, pp. 2455-2460, 2009 IEEE International Conference on Robotics and Automation, ICRA '09, Kobe, Japan, 5/12/09. https://doi.org/10.1109/ROBOT.2009.5152445
Thanh TN, Nagahara H, Sagawa R, Mukaigawa Y, Yachida M, Yagi Y. An adaptive-scale robust estimator for motion estimation. In 2009 IEEE International Conference on Robotics and Automation, ICRA '09. 2009. p. 2455-2460. 5152445 https://doi.org/10.1109/ROBOT.2009.5152445
Thanh, Trung Ngo ; Nagahara, Hajime ; Sagawa, Ryusuke ; Mukaigawa, Yasuhiro ; Yachida, Masahiko ; Yagi, Yasushi. / An adaptive-scale robust estimator for motion estimation. 2009 IEEE International Conference on Robotics and Automation, ICRA '09. 2009. pp. 2455-2460
@inproceedings{209e5e753d3e44ea88dedf84b1adfcb0,
title = "An adaptive-scale robust estimator for motion estimation",
abstract = "Although RANSAC is the most widely used robust estimator in computer vision, it has certain limitations making it ineffective in some situations, such as the motion estimation problem, in which uncertainty on the image features changes according to the capturing conditions. The greatest problem is that the threshold used by RANSAC to detect inliers cannot be changed adaptively; instead it is fixed by the user. An adaptive scale algorithm must therefore be applied in such cases. In this paper, we propose a new adaptive scale robust estimator that adaptively finds the best solution with the best scale to fit the inliers, without the need for predefined information. Our new adaptive scale estimator matches the residual probability density from an estimate and the standard Gaussian probability density function to find the best inlier scale. Our algorithm is evaluated in several motion estimation experiments under varying conditions and the results are compared with several of the latest adaptive-scale robust estimators.",
author = "Thanh, {Trung Ngo} and Hajime Nagahara and Ryusuke Sagawa and Yasuhiro Mukaigawa and Masahiko Yachida and Yasushi Yagi",
year = "2009",
doi = "10.1109/ROBOT.2009.5152445",
language = "English",
isbn = "9781424427895",
pages = "2455--2460",
booktitle = "2009 IEEE International Conference on Robotics and Automation, ICRA '09",

}

TY - GEN

T1 - An adaptive-scale robust estimator for motion estimation

AU - Thanh, Trung Ngo

AU - Nagahara, Hajime

AU - Sagawa, Ryusuke

AU - Mukaigawa, Yasuhiro

AU - Yachida, Masahiko

AU - Yagi, Yasushi

PY - 2009

Y1 - 2009

N2 - Although RANSAC is the most widely used robust estimator in computer vision, it has certain limitations making it ineffective in some situations, such as the motion estimation problem, in which uncertainty on the image features changes according to the capturing conditions. The greatest problem is that the threshold used by RANSAC to detect inliers cannot be changed adaptively; instead it is fixed by the user. An adaptive scale algorithm must therefore be applied in such cases. In this paper, we propose a new adaptive scale robust estimator that adaptively finds the best solution with the best scale to fit the inliers, without the need for predefined information. Our new adaptive scale estimator matches the residual probability density from an estimate and the standard Gaussian probability density function to find the best inlier scale. Our algorithm is evaluated in several motion estimation experiments under varying conditions and the results are compared with several of the latest adaptive-scale robust estimators.

AB - Although RANSAC is the most widely used robust estimator in computer vision, it has certain limitations making it ineffective in some situations, such as the motion estimation problem, in which uncertainty on the image features changes according to the capturing conditions. The greatest problem is that the threshold used by RANSAC to detect inliers cannot be changed adaptively; instead it is fixed by the user. An adaptive scale algorithm must therefore be applied in such cases. In this paper, we propose a new adaptive scale robust estimator that adaptively finds the best solution with the best scale to fit the inliers, without the need for predefined information. Our new adaptive scale estimator matches the residual probability density from an estimate and the standard Gaussian probability density function to find the best inlier scale. Our algorithm is evaluated in several motion estimation experiments under varying conditions and the results are compared with several of the latest adaptive-scale robust estimators.

UR - http://www.scopus.com/inward/record.url?scp=70350362581&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350362581&partnerID=8YFLogxK

U2 - 10.1109/ROBOT.2009.5152445

DO - 10.1109/ROBOT.2009.5152445

M3 - Conference contribution

SN - 9781424427895

SP - 2455

EP - 2460

BT - 2009 IEEE International Conference on Robotics and Automation, ICRA '09

ER -