Robust alignment for panoramic stitching via an exact rank constraint

Yuelong Li, Mohammad Tofighi, Vishal Monga

Research output: Contribution to journalArticle

Abstract

We study the problem of image alignment for panoramic stitching. Unlike most existing approaches that are feature-based, our algorithm works on pixels directly, and accounts for errors across the whole images globally. Technically, we formulate the alignment problem as rank-1 and sparse matrix decomposition over transformed images, and develop an efficient algorithm for solving this challenging non-convex optimization problem. The algorithm reduces to solving a sequence of subproblems, where we analytically establish exact recovery conditions, convergence and optimality, together with convergence rate and complexity. We generalize it to simultaneously align multiple images and recover multiple homographies, extending its application scope toward vast majority of practical scenarios. The experimental results demonstrate that the proposed algorithm is capable of more accurately aligning the images and generating higher quality stitched images than the state-of-the-art methods.

Original languageEnglish (US)
Article number8684316
Pages (from-to)4730-4745
Number of pages16
JournalIEEE Transactions on Image Processing
Volume28
Issue number10
DOIs
StatePublished - Oct 2019

Fingerprint

Image quality
Pixels
Decomposition
Recovery

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Graphics and Computer-Aided Design

Cite this

@article{b4352e1e5d68468c93eac45e58c379eb,
title = "Robust alignment for panoramic stitching via an exact rank constraint",
abstract = "We study the problem of image alignment for panoramic stitching. Unlike most existing approaches that are feature-based, our algorithm works on pixels directly, and accounts for errors across the whole images globally. Technically, we formulate the alignment problem as rank-1 and sparse matrix decomposition over transformed images, and develop an efficient algorithm for solving this challenging non-convex optimization problem. The algorithm reduces to solving a sequence of subproblems, where we analytically establish exact recovery conditions, convergence and optimality, together with convergence rate and complexity. We generalize it to simultaneously align multiple images and recover multiple homographies, extending its application scope toward vast majority of practical scenarios. The experimental results demonstrate that the proposed algorithm is capable of more accurately aligning the images and generating higher quality stitched images than the state-of-the-art methods.",
author = "Yuelong Li and Mohammad Tofighi and Vishal Monga",
year = "2019",
month = "10",
doi = "10.1109/TIP.2019.2909800",
language = "English (US)",
volume = "28",
pages = "4730--4745",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "10",

}

Robust alignment for panoramic stitching via an exact rank constraint. / Li, Yuelong; Tofighi, Mohammad; Monga, Vishal.

In: IEEE Transactions on Image Processing, Vol. 28, No. 10, 8684316, 10.2019, p. 4730-4745.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Robust alignment for panoramic stitching via an exact rank constraint

AU - Li, Yuelong

AU - Tofighi, Mohammad

AU - Monga, Vishal

PY - 2019/10

Y1 - 2019/10

N2 - We study the problem of image alignment for panoramic stitching. Unlike most existing approaches that are feature-based, our algorithm works on pixels directly, and accounts for errors across the whole images globally. Technically, we formulate the alignment problem as rank-1 and sparse matrix decomposition over transformed images, and develop an efficient algorithm for solving this challenging non-convex optimization problem. The algorithm reduces to solving a sequence of subproblems, where we analytically establish exact recovery conditions, convergence and optimality, together with convergence rate and complexity. We generalize it to simultaneously align multiple images and recover multiple homographies, extending its application scope toward vast majority of practical scenarios. The experimental results demonstrate that the proposed algorithm is capable of more accurately aligning the images and generating higher quality stitched images than the state-of-the-art methods.

AB - We study the problem of image alignment for panoramic stitching. Unlike most existing approaches that are feature-based, our algorithm works on pixels directly, and accounts for errors across the whole images globally. Technically, we formulate the alignment problem as rank-1 and sparse matrix decomposition over transformed images, and develop an efficient algorithm for solving this challenging non-convex optimization problem. The algorithm reduces to solving a sequence of subproblems, where we analytically establish exact recovery conditions, convergence and optimality, together with convergence rate and complexity. We generalize it to simultaneously align multiple images and recover multiple homographies, extending its application scope toward vast majority of practical scenarios. The experimental results demonstrate that the proposed algorithm is capable of more accurately aligning the images and generating higher quality stitched images than the state-of-the-art methods.

UR - http://www.scopus.com/inward/record.url?scp=85070472891&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070472891&partnerID=8YFLogxK

U2 - 10.1109/TIP.2019.2909800

DO - 10.1109/TIP.2019.2909800

M3 - Article

C2 - 30969922

AN - SCOPUS:85070472891

VL - 28

SP - 4730

EP - 4745

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 10

M1 - 8684316

ER -