Robust image retargeting via axis-aligned deformation

Daniele Panozzo, Ofir Weber, Olga Sorkine

Research output: Contribution to journalArticle

Abstract

We propose the space of axis-aligned deformations as the meaningful space for content-aware image retargeting. Such deformations exclude local rotations, avoiding harmful visual distortions, and they are parameterized in 1D. We show that standard warping energies for image retargeting can be minimized in the space of axis-aligned deformations while guaranteeing that bijectivity constraints are satisfied, leading to high-quality, smooth and robust retargeting results. Thanks to the 1D parameterization, our method only requires solving a small quadratic program, which can be done within a few milliseconds on the CPU with no precomputation overhead. We demonstrate how the image size and the saliency map can be changed in real time with our approach, and present results on various input images, including the RETARGETME benchmark. We compare our results with six other algorithms in a user study to demonstrate that the space of axis-aligned deformations is suitable for the problem at hand.

Original languageEnglish (US)
Pages (from-to)229-236
Number of pages8
JournalComputer Graphics Forum
Volume31
Issue number2
DOIs
StatePublished - 2012

Fingerprint

Parameterization
Program processors

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Robust image retargeting via axis-aligned deformation. / Panozzo, Daniele; Weber, Ofir; Sorkine, Olga.

In: Computer Graphics Forum, Vol. 31, No. 2, 2012, p. 229-236.

Research output: Contribution to journalArticle

Panozzo, Daniele ; Weber, Ofir ; Sorkine, Olga. / Robust image retargeting via axis-aligned deformation. In: Computer Graphics Forum. 2012 ; Vol. 31, No. 2. pp. 229-236.
@article{3f27b6e4504c4caba10653a831d2171a,
title = "Robust image retargeting via axis-aligned deformation",
abstract = "We propose the space of axis-aligned deformations as the meaningful space for content-aware image retargeting. Such deformations exclude local rotations, avoiding harmful visual distortions, and they are parameterized in 1D. We show that standard warping energies for image retargeting can be minimized in the space of axis-aligned deformations while guaranteeing that bijectivity constraints are satisfied, leading to high-quality, smooth and robust retargeting results. Thanks to the 1D parameterization, our method only requires solving a small quadratic program, which can be done within a few milliseconds on the CPU with no precomputation overhead. We demonstrate how the image size and the saliency map can be changed in real time with our approach, and present results on various input images, including the RETARGETME benchmark. We compare our results with six other algorithms in a user study to demonstrate that the space of axis-aligned deformations is suitable for the problem at hand.",
author = "Daniele Panozzo and Ofir Weber and Olga Sorkine",
year = "2012",
doi = "10.1111/j.1467-8659.2012.03001.x",
language = "English (US)",
volume = "31",
pages = "229--236",
journal = "Computer Graphics Forum",
issn = "0167-7055",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - Robust image retargeting via axis-aligned deformation

AU - Panozzo, Daniele

AU - Weber, Ofir

AU - Sorkine, Olga

PY - 2012

Y1 - 2012

N2 - We propose the space of axis-aligned deformations as the meaningful space for content-aware image retargeting. Such deformations exclude local rotations, avoiding harmful visual distortions, and they are parameterized in 1D. We show that standard warping energies for image retargeting can be minimized in the space of axis-aligned deformations while guaranteeing that bijectivity constraints are satisfied, leading to high-quality, smooth and robust retargeting results. Thanks to the 1D parameterization, our method only requires solving a small quadratic program, which can be done within a few milliseconds on the CPU with no precomputation overhead. We demonstrate how the image size and the saliency map can be changed in real time with our approach, and present results on various input images, including the RETARGETME benchmark. We compare our results with six other algorithms in a user study to demonstrate that the space of axis-aligned deformations is suitable for the problem at hand.

AB - We propose the space of axis-aligned deformations as the meaningful space for content-aware image retargeting. Such deformations exclude local rotations, avoiding harmful visual distortions, and they are parameterized in 1D. We show that standard warping energies for image retargeting can be minimized in the space of axis-aligned deformations while guaranteeing that bijectivity constraints are satisfied, leading to high-quality, smooth and robust retargeting results. Thanks to the 1D parameterization, our method only requires solving a small quadratic program, which can be done within a few milliseconds on the CPU with no precomputation overhead. We demonstrate how the image size and the saliency map can be changed in real time with our approach, and present results on various input images, including the RETARGETME benchmark. We compare our results with six other algorithms in a user study to demonstrate that the space of axis-aligned deformations is suitable for the problem at hand.

UR - http://www.scopus.com/inward/record.url?scp=84877619224&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84877619224&partnerID=8YFLogxK

U2 - 10.1111/j.1467-8659.2012.03001.x

DO - 10.1111/j.1467-8659.2012.03001.x

M3 - Article

VL - 31

SP - 229

EP - 236

JO - Computer Graphics Forum

JF - Computer Graphics Forum

SN - 0167-7055

IS - 2

ER -