Disparity estimation using a region-dividing technique and energy-based regularization

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

A two-stage algorithm is proposed for locating smooth and detailed disparity vector fields in a stereo image pair. The algorithm consists of hierarchical disparity estimation using a region-dividing technique and edge-preserving regularization. The hierarchical region-dividing disparity estimation increases the efficiency and reliability of the estimation process. At the second stage, the vector fields are regularized with an energy model that produces smooth fields while preserving discontinuities resulting from object boundaries. The minimization problem is addressed by solving a corresponding partial differential equation using a finite-difference method. Experiments show that the proposed algorithm provides accurate and spatially correlated dispartly vector fields in various types of stereo images, even in the case of images with large displacements.

Original languageEnglish
Pages (from-to)1882-1890
Number of pages9
JournalOptical Engineering
Volume43
Issue number8
DOIs
Publication statusPublished - 2004 Aug 1

Fingerprint

preserving
Finite difference method
partial differential equations
Partial differential equations
energy
discontinuity
optimization
Experiments

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics
  • Engineering(all)

Cite this

@article{db2717f63a26464794f381dc4de10bdc,
title = "Disparity estimation using a region-dividing technique and energy-based regularization",
abstract = "A two-stage algorithm is proposed for locating smooth and detailed disparity vector fields in a stereo image pair. The algorithm consists of hierarchical disparity estimation using a region-dividing technique and edge-preserving regularization. The hierarchical region-dividing disparity estimation increases the efficiency and reliability of the estimation process. At the second stage, the vector fields are regularized with an energy model that produces smooth fields while preserving discontinuities resulting from object boundaries. The minimization problem is addressed by solving a corresponding partial differential equation using a finite-difference method. Experiments show that the proposed algorithm provides accurate and spatially correlated dispartly vector fields in various types of stereo images, even in the case of images with large displacements.",
author = "Hansung Kim and Yoonsik Choe and Kwanghoon Sohn",
year = "2004",
month = "8",
day = "1",
doi = "10.1117/1.1766299",
language = "English",
volume = "43",
pages = "1882--1890",
journal = "Optical Engineering",
issn = "0091-3286",
publisher = "SPIE",
number = "8",

}

Disparity estimation using a region-dividing technique and energy-based regularization. / Kim, Hansung; Choe, Yoonsik; Sohn, Kwanghoon.

In: Optical Engineering, Vol. 43, No. 8, 01.08.2004, p. 1882-1890.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Disparity estimation using a region-dividing technique and energy-based regularization

AU - Kim, Hansung

AU - Choe, Yoonsik

AU - Sohn, Kwanghoon

PY - 2004/8/1

Y1 - 2004/8/1

N2 - A two-stage algorithm is proposed for locating smooth and detailed disparity vector fields in a stereo image pair. The algorithm consists of hierarchical disparity estimation using a region-dividing technique and edge-preserving regularization. The hierarchical region-dividing disparity estimation increases the efficiency and reliability of the estimation process. At the second stage, the vector fields are regularized with an energy model that produces smooth fields while preserving discontinuities resulting from object boundaries. The minimization problem is addressed by solving a corresponding partial differential equation using a finite-difference method. Experiments show that the proposed algorithm provides accurate and spatially correlated dispartly vector fields in various types of stereo images, even in the case of images with large displacements.

AB - A two-stage algorithm is proposed for locating smooth and detailed disparity vector fields in a stereo image pair. The algorithm consists of hierarchical disparity estimation using a region-dividing technique and edge-preserving regularization. The hierarchical region-dividing disparity estimation increases the efficiency and reliability of the estimation process. At the second stage, the vector fields are regularized with an energy model that produces smooth fields while preserving discontinuities resulting from object boundaries. The minimization problem is addressed by solving a corresponding partial differential equation using a finite-difference method. Experiments show that the proposed algorithm provides accurate and spatially correlated dispartly vector fields in various types of stereo images, even in the case of images with large displacements.

UR - http://www.scopus.com/inward/record.url?scp=4644268711&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=4644268711&partnerID=8YFLogxK

U2 - 10.1117/1.1766299

DO - 10.1117/1.1766299

M3 - Article

AN - SCOPUS:4644268711

VL - 43

SP - 1882

EP - 1890

JO - Optical Engineering

JF - Optical Engineering

SN - 0091-3286

IS - 8

ER -