Globally consistent correspondence of multiple feature sets using proximal Gauss-Seidel relaxation

Jin Gang Yu, Gui Song Xia, Ashok K Samal, Jinwen Tian

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Feature correspondence between two or more images is a fundamental problem towards many computer vision applications. The case of correspondence between two images has been intensively studied, however, few works so far have been concerned with multi-image correspondence. In this paper, we address the problem of establishing a globally consistent correspondence among multiple (more than two) feature sets given the pairwise feature affinity information. Our main contribution is to propose a novel optimization framework for solving this problem based on the so-called Proximal Gauss-Seidel Relaxation (PGSR). The proposed method is distinguished from previous works mainly in three aspects: (1) it is more robust to noise and outliers; (2) its solution is based on convex relaxation and the principled PGSR method, which in general has convergence guarantee; (3) the scale of the problem in our method is linear with respect to the number of feature sets, making it computationally practical to be used in real-world applications. Experimental results both synthetic and real image datasets have demonstrated the effectiveness and superiority of the proposed method.

Original languageEnglish (US)
Pages (from-to)255-267
Number of pages13
JournalPattern Recognition
Volume51
DOIs
StatePublished - Mar 1 2016

Fingerprint

Computer vision

Keywords

  • Convex relaxation
  • Feature correspondence
  • Graph matching
  • Multiple feature set correspondence
  • Permutation matrix
  • Proximal Gauss-Seidel method

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

Globally consistent correspondence of multiple feature sets using proximal Gauss-Seidel relaxation. / Yu, Jin Gang; Xia, Gui Song; Samal, Ashok K; Tian, Jinwen.

In: Pattern Recognition, Vol. 51, 01.03.2016, p. 255-267.

Research output: Contribution to journalArticle

@article{144c3d5833f9438685366b4438c3d171,
title = "Globally consistent correspondence of multiple feature sets using proximal Gauss-Seidel relaxation",
abstract = "Feature correspondence between two or more images is a fundamental problem towards many computer vision applications. The case of correspondence between two images has been intensively studied, however, few works so far have been concerned with multi-image correspondence. In this paper, we address the problem of establishing a globally consistent correspondence among multiple (more than two) feature sets given the pairwise feature affinity information. Our main contribution is to propose a novel optimization framework for solving this problem based on the so-called Proximal Gauss-Seidel Relaxation (PGSR). The proposed method is distinguished from previous works mainly in three aspects: (1) it is more robust to noise and outliers; (2) its solution is based on convex relaxation and the principled PGSR method, which in general has convergence guarantee; (3) the scale of the problem in our method is linear with respect to the number of feature sets, making it computationally practical to be used in real-world applications. Experimental results both synthetic and real image datasets have demonstrated the effectiveness and superiority of the proposed method.",
keywords = "Convex relaxation, Feature correspondence, Graph matching, Multiple feature set correspondence, Permutation matrix, Proximal Gauss-Seidel method",
author = "Yu, {Jin Gang} and Xia, {Gui Song} and Samal, {Ashok K} and Jinwen Tian",
year = "2016",
month = "3",
day = "1",
doi = "10.1016/j.patcog.2015.09.029",
language = "English (US)",
volume = "51",
pages = "255--267",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Globally consistent correspondence of multiple feature sets using proximal Gauss-Seidel relaxation

AU - Yu, Jin Gang

AU - Xia, Gui Song

AU - Samal, Ashok K

AU - Tian, Jinwen

PY - 2016/3/1

Y1 - 2016/3/1

N2 - Feature correspondence between two or more images is a fundamental problem towards many computer vision applications. The case of correspondence between two images has been intensively studied, however, few works so far have been concerned with multi-image correspondence. In this paper, we address the problem of establishing a globally consistent correspondence among multiple (more than two) feature sets given the pairwise feature affinity information. Our main contribution is to propose a novel optimization framework for solving this problem based on the so-called Proximal Gauss-Seidel Relaxation (PGSR). The proposed method is distinguished from previous works mainly in three aspects: (1) it is more robust to noise and outliers; (2) its solution is based on convex relaxation and the principled PGSR method, which in general has convergence guarantee; (3) the scale of the problem in our method is linear with respect to the number of feature sets, making it computationally practical to be used in real-world applications. Experimental results both synthetic and real image datasets have demonstrated the effectiveness and superiority of the proposed method.

AB - Feature correspondence between two or more images is a fundamental problem towards many computer vision applications. The case of correspondence between two images has been intensively studied, however, few works so far have been concerned with multi-image correspondence. In this paper, we address the problem of establishing a globally consistent correspondence among multiple (more than two) feature sets given the pairwise feature affinity information. Our main contribution is to propose a novel optimization framework for solving this problem based on the so-called Proximal Gauss-Seidel Relaxation (PGSR). The proposed method is distinguished from previous works mainly in three aspects: (1) it is more robust to noise and outliers; (2) its solution is based on convex relaxation and the principled PGSR method, which in general has convergence guarantee; (3) the scale of the problem in our method is linear with respect to the number of feature sets, making it computationally practical to be used in real-world applications. Experimental results both synthetic and real image datasets have demonstrated the effectiveness and superiority of the proposed method.

KW - Convex relaxation

KW - Feature correspondence

KW - Graph matching

KW - Multiple feature set correspondence

KW - Permutation matrix

KW - Proximal Gauss-Seidel method

UR - http://www.scopus.com/inward/record.url?scp=84955703716&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955703716&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2015.09.029

DO - 10.1016/j.patcog.2015.09.029

M3 - Article

AN - SCOPUS:84955703716

VL - 51

SP - 255

EP - 267

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

ER -