Mahalanobis distance cross-correlation for illumination-invariant stereo matching

Seungryong Kim, Bumsub Ham, Bongjoe Kim, Kwanghoon Sohn

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

A robust similarity measure called the Mahalanobis distance cross-correlation (MDCC) is proposed for illumination-invariant stereo matching, which uses a local color distribution within support windows. It is shown that the Mahalanobis distance between the color itself and the average color is preserved under affine transformation. The MDCC converts pixels within each support window into the Mahalanobis distance transform (MDT) space. The similarity between MDT pairs is then computed using the cross-correlation with an asymmetric weight function based on the Mahalanobis distance. The MDCC considers correlation on cross-color channels, thus providing robustness to affine illumination variation. Experimental results show that the MDCC outperforms state-of-the-art similarity measures in terms of stereo matching for image pairs taken under different illumination conditions.

Original languageEnglish
Article number6826490
Pages (from-to)1844-1859
Number of pages16
JournalIEEE Transactions on Circuits and Systems for Video Technology
Volume24
Issue number11
DOIs
Publication statusPublished - 2014 Nov 1

Fingerprint

Lighting
Color
Pixels

All Science Journal Classification (ASJC) codes

  • Media Technology
  • Electrical and Electronic Engineering

Cite this

@article{ad6f8260b64a400d954e42e6a0e0aa1b,
title = "Mahalanobis distance cross-correlation for illumination-invariant stereo matching",
abstract = "A robust similarity measure called the Mahalanobis distance cross-correlation (MDCC) is proposed for illumination-invariant stereo matching, which uses a local color distribution within support windows. It is shown that the Mahalanobis distance between the color itself and the average color is preserved under affine transformation. The MDCC converts pixels within each support window into the Mahalanobis distance transform (MDT) space. The similarity between MDT pairs is then computed using the cross-correlation with an asymmetric weight function based on the Mahalanobis distance. The MDCC considers correlation on cross-color channels, thus providing robustness to affine illumination variation. Experimental results show that the MDCC outperforms state-of-the-art similarity measures in terms of stereo matching for image pairs taken under different illumination conditions.",
author = "Seungryong Kim and Bumsub Ham and Bongjoe Kim and Kwanghoon Sohn",
year = "2014",
month = "11",
day = "1",
doi = "10.1109/TCSVT.2014.2329377",
language = "English",
volume = "24",
pages = "1844--1859",
journal = "IEEE Transactions on Circuits and Systems for Video Technology",
issn = "1051-8215",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "11",

}

Mahalanobis distance cross-correlation for illumination-invariant stereo matching. / Kim, Seungryong; Ham, Bumsub; Kim, Bongjoe; Sohn, Kwanghoon.

In: IEEE Transactions on Circuits and Systems for Video Technology, Vol. 24, No. 11, 6826490, 01.11.2014, p. 1844-1859.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Mahalanobis distance cross-correlation for illumination-invariant stereo matching

AU - Kim, Seungryong

AU - Ham, Bumsub

AU - Kim, Bongjoe

AU - Sohn, Kwanghoon

PY - 2014/11/1

Y1 - 2014/11/1

N2 - A robust similarity measure called the Mahalanobis distance cross-correlation (MDCC) is proposed for illumination-invariant stereo matching, which uses a local color distribution within support windows. It is shown that the Mahalanobis distance between the color itself and the average color is preserved under affine transformation. The MDCC converts pixels within each support window into the Mahalanobis distance transform (MDT) space. The similarity between MDT pairs is then computed using the cross-correlation with an asymmetric weight function based on the Mahalanobis distance. The MDCC considers correlation on cross-color channels, thus providing robustness to affine illumination variation. Experimental results show that the MDCC outperforms state-of-the-art similarity measures in terms of stereo matching for image pairs taken under different illumination conditions.

AB - A robust similarity measure called the Mahalanobis distance cross-correlation (MDCC) is proposed for illumination-invariant stereo matching, which uses a local color distribution within support windows. It is shown that the Mahalanobis distance between the color itself and the average color is preserved under affine transformation. The MDCC converts pixels within each support window into the Mahalanobis distance transform (MDT) space. The similarity between MDT pairs is then computed using the cross-correlation with an asymmetric weight function based on the Mahalanobis distance. The MDCC considers correlation on cross-color channels, thus providing robustness to affine illumination variation. Experimental results show that the MDCC outperforms state-of-the-art similarity measures in terms of stereo matching for image pairs taken under different illumination conditions.

UR - http://www.scopus.com/inward/record.url?scp=84908655258&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908655258&partnerID=8YFLogxK

U2 - 10.1109/TCSVT.2014.2329377

DO - 10.1109/TCSVT.2014.2329377

M3 - Article

VL - 24

SP - 1844

EP - 1859

JO - IEEE Transactions on Circuits and Systems for Video Technology

JF - IEEE Transactions on Circuits and Systems for Video Technology

SN - 1051-8215

IS - 11

M1 - 6826490

ER -