Fast inverse compositional image alignment with missing data and re-weighting

Vincent Wen Han Lui; Dinesh Srikantha Gamage; Thomas William Drummond

doi:10.5244/C.29.54

Fast inverse compositional image alignment with missing data and re-weighting

Vincent Wen Han Lui, Dinesh Srikantha Gamage, Thomas William Drummond

Electrical and Computer Systems Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

Abstract

This paper proposes a novel method of performing inverse compositional image
alignment which elegantly deals with missing data and re-weighting, and does not require the Jacobians and Hessian to be re-computed at every iteration. We show how missing data and re-weighting can be handled through preconditioning. We propose a few preconditioning techniques and analyse how each technique models the effects of missing data and re-weighting for inverse composition. We show through extensive experiments on different applications that our method improves the convergence rate of the conventional re-weighted inverse compositional method while remaining robust to outliers. We also show that the update parameters are usually underestimated and how this can be used to further speed up convergence of image alignment methods.

Original language	English
Title of host publication	Proceedings of the British Machine Vision Conference 2015
Editors	Xianghua Xie, Mark W Jones, Gary K L Tam
Place of Publication	Durham UK
Publisher	British Machine Vision Association
Pages	1 - 12
Number of pages	12
ISBN (Print)	1901725537
DOIs	https://doi.org/10.5244/C.29.54
Publication status	Published - 2015
Event	British Machine Vision Conference 2015 - Swansea, United Kingdom Duration: 7 Sept 2015 → 10 Sept 2015 Conference number: 26th http://www.bmva.org/bmvc/2015/index.html https://dblp.org/db/conf/bmvc/bmvc2015.html

Conference

Conference	British Machine Vision Conference 2015
Abbreviated title	BMVC 2015
Country/Territory	United Kingdom
City	Swansea
Period	7/09/15 → 10/09/15
Internet address	http://www.bmva.org/bmvc/2015/index.html https://dblp.org/db/conf/bmvc/bmvc2015.html

Access to Document

10.5244/C.29.54

Cite this

@inproceedings{4596d3030a7c44a2a8bf1f9d5f355984,

title = "Fast inverse compositional image alignment with missing data and re-weighting",

abstract = "This paper proposes a novel method of performing inverse compositional imagealignment which elegantly deals with missing data and re-weighting, and does not require the Jacobians and Hessian to be re-computed at every iteration. We show how missing data and re-weighting can be handled through preconditioning. We propose a few preconditioning techniques and analyse how each technique models the effects of missing data and re-weighting for inverse composition. We show through extensive experiments on different applications that our method improves the convergence rate of the conventional re-weighted inverse compositional method while remaining robust to outliers. We also show that the update parameters are usually underestimated and how this can be used to further speed up convergence of image alignment methods.",

author = "Lui, {Vincent Wen Han} and Gamage, {Dinesh Srikantha} and Drummond, {Thomas William}",

year = "2015",

doi = "10.5244/C.29.54",

language = "English",

isbn = "1901725537",

pages = "1 -- 12",

editor = "Xianghua Xie and Jones, {Mark W} and Tam, {Gary K L}",

booktitle = "Proceedings of the British Machine Vision Conference 2015",

publisher = "British Machine Vision Association",

note = "British Machine Vision Conference 2015, BMVC 2015 ; Conference date: 07-09-2015 Through 10-09-2015",

url = "http://www.bmva.org/bmvc/2015/index.html, https://dblp.org/db/conf/bmvc/bmvc2015.html",

}

Lui, VWH, Gamage, DS & Drummond, TW 2015, Fast inverse compositional image alignment with missing data and re-weighting. in X Xie, MW Jones & GKL Tam (eds), Proceedings of the British Machine Vision Conference 2015. British Machine Vision Association, Durham UK, pp. 1 - 12, British Machine Vision Conference 2015, Swansea, United Kingdom, 7/09/15. https://doi.org/10.5244/C.29.54

Fast inverse compositional image alignment with missing data and re-weighting. / Lui, Vincent Wen Han; Gamage, Dinesh Srikantha; Drummond, Thomas William.
Proceedings of the British Machine Vision Conference 2015. ed. / Xianghua Xie; Mark W Jones; Gary K L Tam. Durham UK: British Machine Vision Association, 2015. p. 1 - 12.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Fast inverse compositional image alignment with missing data and re-weighting

AU - Lui, Vincent Wen Han

AU - Gamage, Dinesh Srikantha

AU - Drummond, Thomas William

N1 - Conference code: 26th

PY - 2015

Y1 - 2015

N2 - This paper proposes a novel method of performing inverse compositional imagealignment which elegantly deals with missing data and re-weighting, and does not require the Jacobians and Hessian to be re-computed at every iteration. We show how missing data and re-weighting can be handled through preconditioning. We propose a few preconditioning techniques and analyse how each technique models the effects of missing data and re-weighting for inverse composition. We show through extensive experiments on different applications that our method improves the convergence rate of the conventional re-weighted inverse compositional method while remaining robust to outliers. We also show that the update parameters are usually underestimated and how this can be used to further speed up convergence of image alignment methods.

AB - This paper proposes a novel method of performing inverse compositional imagealignment which elegantly deals with missing data and re-weighting, and does not require the Jacobians and Hessian to be re-computed at every iteration. We show how missing data and re-weighting can be handled through preconditioning. We propose a few preconditioning techniques and analyse how each technique models the effects of missing data and re-weighting for inverse composition. We show through extensive experiments on different applications that our method improves the convergence rate of the conventional re-weighted inverse compositional method while remaining robust to outliers. We also show that the update parameters are usually underestimated and how this can be used to further speed up convergence of image alignment methods.

UR - http://www.bmva.org/bmvc/2015/papers/paper054/paper054.pdf

U2 - 10.5244/C.29.54

DO - 10.5244/C.29.54

M3 - Conference Paper

SN - 1901725537

SP - 1

EP - 12

BT - Proceedings of the British Machine Vision Conference 2015

A2 - Xie, Xianghua

A2 - Jones, Mark W

A2 - Tam, Gary K L

PB - British Machine Vision Association

CY - Durham UK

T2 - British Machine Vision Conference 2015

Y2 - 7 September 2015 through 10 September 2015

ER -

Fast inverse compositional image alignment with missing data and re-weighting

Abstract

Conference

Access to Document

Other files and links

Cite this