Using genetic algorithms in computer vision: registering images to 3D surface model

Zsolt Jankó*, Dmitry Chetverikov, Anikó Ekárt

*Corresponding author for this work

Research output: Contribution to journalArticle

Abstract

This paper shows a successful application of genetic algorithms in computer vision. We aim at building photorealistic 3D models of real-world objects by adding textural information to the geometry. In this paper we focus on the 2D-3D registration problem: given a 3D geometric model of an object, and optical images of the same object, we need to find the precise alignment of the 2D images to the 3D model. We generalise the photo-consistency approach of Clarkson et al. who assume calibrated cameras, thus only the pose of the object in the world needs to be estimated. Our method extends this approach to the case of uncalibrated cameras, when both intrinsic and extrinsic camera parameters are unknown. We formulate the problem as an optimisation and use a genetic algorithm to find a solution. We use semi-synthetic data to study the effects of different parameter settings on the registration. Additionally, experimental results on real data are presented to demonstrate the efficiency of the method.

Original languageEnglish
Pages (from-to)193-212
Number of pages20
JournalUniversity of Szeged. Acta Cybernetica
Volume18
Issue number2
Publication statusPublished - 1 Jan 2007

Fingerprint

Computer Vision
Computer vision
Genetic algorithms
Cameras
3D Model
Genetic Algorithm
Camera
Registration
Geometric Model
Synthetic Data
Model
Alignment
Geometry
Unknown
Generalise
Object
Genetic algorithm
Optimization
Experimental Results
Demonstrate

Keywords

  • Photo-consistency
  • Photorealistic models
  • Uncalibrated images

Cite this

@article{5a56f63069d44f12963f54518107d6a4,
title = "Using genetic algorithms in computer vision: registering images to 3D surface model",
abstract = "This paper shows a successful application of genetic algorithms in computer vision. We aim at building photorealistic 3D models of real-world objects by adding textural information to the geometry. In this paper we focus on the 2D-3D registration problem: given a 3D geometric model of an object, and optical images of the same object, we need to find the precise alignment of the 2D images to the 3D model. We generalise the photo-consistency approach of Clarkson et al. who assume calibrated cameras, thus only the pose of the object in the world needs to be estimated. Our method extends this approach to the case of uncalibrated cameras, when both intrinsic and extrinsic camera parameters are unknown. We formulate the problem as an optimisation and use a genetic algorithm to find a solution. We use semi-synthetic data to study the effects of different parameter settings on the registration. Additionally, experimental results on real data are presented to demonstrate the efficiency of the method.",
keywords = "Photo-consistency, Photorealistic models, Uncalibrated images",
author = "Zsolt Jank{\'o} and Dmitry Chetverikov and Anik{\'o} Ek{\'a}rt",
year = "2007",
month = "1",
day = "1",
language = "English",
volume = "18",
pages = "193--212",
journal = "University of Szeged. Acta Cybernetica",
issn = "0324-721X",
publisher = "University of Szeged",
number = "2",

}

Using genetic algorithms in computer vision : registering images to 3D surface model. / Jankó, Zsolt; Chetverikov, Dmitry; Ekárt, Anikó.

In: University of Szeged. Acta Cybernetica, Vol. 18, No. 2, 01.01.2007, p. 193-212.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Using genetic algorithms in computer vision

T2 - registering images to 3D surface model

AU - Jankó, Zsolt

AU - Chetverikov, Dmitry

AU - Ekárt, Anikó

PY - 2007/1/1

Y1 - 2007/1/1

N2 - This paper shows a successful application of genetic algorithms in computer vision. We aim at building photorealistic 3D models of real-world objects by adding textural information to the geometry. In this paper we focus on the 2D-3D registration problem: given a 3D geometric model of an object, and optical images of the same object, we need to find the precise alignment of the 2D images to the 3D model. We generalise the photo-consistency approach of Clarkson et al. who assume calibrated cameras, thus only the pose of the object in the world needs to be estimated. Our method extends this approach to the case of uncalibrated cameras, when both intrinsic and extrinsic camera parameters are unknown. We formulate the problem as an optimisation and use a genetic algorithm to find a solution. We use semi-synthetic data to study the effects of different parameter settings on the registration. Additionally, experimental results on real data are presented to demonstrate the efficiency of the method.

AB - This paper shows a successful application of genetic algorithms in computer vision. We aim at building photorealistic 3D models of real-world objects by adding textural information to the geometry. In this paper we focus on the 2D-3D registration problem: given a 3D geometric model of an object, and optical images of the same object, we need to find the precise alignment of the 2D images to the 3D model. We generalise the photo-consistency approach of Clarkson et al. who assume calibrated cameras, thus only the pose of the object in the world needs to be estimated. Our method extends this approach to the case of uncalibrated cameras, when both intrinsic and extrinsic camera parameters are unknown. We formulate the problem as an optimisation and use a genetic algorithm to find a solution. We use semi-synthetic data to study the effects of different parameter settings on the registration. Additionally, experimental results on real data are presented to demonstrate the efficiency of the method.

KW - Photo-consistency

KW - Photorealistic models

KW - Uncalibrated images

UR - http://www.scopus.com/inward/record.url?scp=38049151004&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:38049151004

VL - 18

SP - 193

EP - 212

JO - University of Szeged. Acta Cybernetica

JF - University of Szeged. Acta Cybernetica

SN - 0324-721X

IS - 2

ER -