Realigning 2D and 3D Object Fragments without Correspondences

Csaba Domokos, Z. Kato

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This paper addresses the problem of simultaneous estimation of different linear deformations, resulting in a global non-linear transformation, between an original object and its broken fragments. A general framework is proposed without using correspondences, where the solution of a polynomial system of equations directly provides the parameters of the alignment. We quantitatively evaluate the proposed algorithm on a large synthetic dataset containing 2D and 3D images, where linear (rigid-body and affine) transformations are considered. We also conduct an exhaustive analysis of the robustness against segmentation errors and the numerical stability of the proposed method. Moreover, we present experiments on 2D real images as well as on volumetric medical images.

Original languageEnglish
Article number7138636
Pages (from-to)195-202
Number of pages8
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume38
Issue number1
DOIs
Publication statusPublished - Jan 1 2016

Fingerprint

Convergence of numerical methods
Fragment
Correspondence
Polynomials
Simultaneous Estimation
Nonlinear Transformation
Polynomial Systems
Numerical Stability
3D Image
Medical Image
Rigid Body
Affine transformation
System of equations
Alignment
Segmentation
Experiments
Robustness
Evaluate
Experiment
Object

Keywords

  • affine puzzle
  • Image registration
  • locally linear deformation
  • realignment of broken fragments

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Software
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

Realigning 2D and 3D Object Fragments without Correspondences. / Domokos, Csaba; Kato, Z.

In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 38, No. 1, 7138636, 01.01.2016, p. 195-202.

Research output: Contribution to journalArticle

@article{e3a67636e7564811a0167220fa1de77d,
title = "Realigning 2D and 3D Object Fragments without Correspondences",
abstract = "This paper addresses the problem of simultaneous estimation of different linear deformations, resulting in a global non-linear transformation, between an original object and its broken fragments. A general framework is proposed without using correspondences, where the solution of a polynomial system of equations directly provides the parameters of the alignment. We quantitatively evaluate the proposed algorithm on a large synthetic dataset containing 2D and 3D images, where linear (rigid-body and affine) transformations are considered. We also conduct an exhaustive analysis of the robustness against segmentation errors and the numerical stability of the proposed method. Moreover, we present experiments on 2D real images as well as on volumetric medical images.",
keywords = "affine puzzle, Image registration, locally linear deformation, realignment of broken fragments",
author = "Csaba Domokos and Z. Kato",
year = "2016",
month = "1",
day = "1",
doi = "10.1109/TPAMI.2015.2450726",
language = "English",
volume = "38",
pages = "195--202",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "1",

}

TY - JOUR

T1 - Realigning 2D and 3D Object Fragments without Correspondences

AU - Domokos, Csaba

AU - Kato, Z.

PY - 2016/1/1

Y1 - 2016/1/1

N2 - This paper addresses the problem of simultaneous estimation of different linear deformations, resulting in a global non-linear transformation, between an original object and its broken fragments. A general framework is proposed without using correspondences, where the solution of a polynomial system of equations directly provides the parameters of the alignment. We quantitatively evaluate the proposed algorithm on a large synthetic dataset containing 2D and 3D images, where linear (rigid-body and affine) transformations are considered. We also conduct an exhaustive analysis of the robustness against segmentation errors and the numerical stability of the proposed method. Moreover, we present experiments on 2D real images as well as on volumetric medical images.

AB - This paper addresses the problem of simultaneous estimation of different linear deformations, resulting in a global non-linear transformation, between an original object and its broken fragments. A general framework is proposed without using correspondences, where the solution of a polynomial system of equations directly provides the parameters of the alignment. We quantitatively evaluate the proposed algorithm on a large synthetic dataset containing 2D and 3D images, where linear (rigid-body and affine) transformations are considered. We also conduct an exhaustive analysis of the robustness against segmentation errors and the numerical stability of the proposed method. Moreover, we present experiments on 2D real images as well as on volumetric medical images.

KW - affine puzzle

KW - Image registration

KW - locally linear deformation

KW - realignment of broken fragments

UR - http://www.scopus.com/inward/record.url?scp=84961572714&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961572714&partnerID=8YFLogxK

U2 - 10.1109/TPAMI.2015.2450726

DO - 10.1109/TPAMI.2015.2450726

M3 - Article

VL - 38

SP - 195

EP - 202

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 1

M1 - 7138636

ER -