Real-time online video object silhouette extraction using graph cuts on the GPU

Zachary A. Garrett, Hideo Saito

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

Being able to find the silhouette of an object is a very important front-end processing step for many high-level computer vision techniques, such as Shape-from-Silhouette 3D reconstruction methods, object shape tracking, and pose estimation. Graph cuts have been proposed as a method for finding very accurate silhouettes which can be used as input to such high level techniques, but graph cuts are notoriously computation intensive and slow. Leading CPU implementations can extract a silhouette from a single QVGA image in 100 milliseconds, with performance dramatically decreasing with increased resolution. Recent GPU implementations have been able to achieve performance of 6 milliseconds per image by exploiting the intrinsic properties of the lattice graphs and the hardware model of the GPU. However, these methods are restricted to a subclass of lattice graphs and are not generally applicable. We propose a novel method for graph cuts on the GPU which places no limits on graph configuration and which is able to achieve comparable real-time performance in online video processing scenarios.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages985-994
Number of pages10
Volume5716 LNCS
DOIs
Publication statusPublished - 2009
Event15th International Conference on Image Analysis and Processing - ICIAP 2009, Proceedings - Vietri sul Mare, Italy
Duration: 2009 Sep 82009 Sep 11

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5716 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other15th International Conference on Image Analysis and Processing - ICIAP 2009, Proceedings
CountryItaly
CityVietri sul Mare
Period09/9/809/9/11

Fingerprint

Graph Cuts
Silhouette
Real-time
Graph in graph theory
Processing
Video Processing
Computer vision
Program processors
Pose Estimation
3D Reconstruction
Computer Vision
Hardware
Scenarios
Configuration
Object
Graphics processing unit

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Garrett, Z. A., & Saito, H. (2009). Real-time online video object silhouette extraction using graph cuts on the GPU. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5716 LNCS, pp. 985-994). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5716 LNCS). https://doi.org/10.1007/978-3-642-04146-4_105

Real-time online video object silhouette extraction using graph cuts on the GPU. / Garrett, Zachary A.; Saito, Hideo.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5716 LNCS 2009. p. 985-994 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5716 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Garrett, ZA & Saito, H 2009, Real-time online video object silhouette extraction using graph cuts on the GPU. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 5716 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5716 LNCS, pp. 985-994, 15th International Conference on Image Analysis and Processing - ICIAP 2009, Proceedings, Vietri sul Mare, Italy, 09/9/8. https://doi.org/10.1007/978-3-642-04146-4_105
Garrett ZA, Saito H. Real-time online video object silhouette extraction using graph cuts on the GPU. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5716 LNCS. 2009. p. 985-994. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-04146-4_105
Garrett, Zachary A. ; Saito, Hideo. / Real-time online video object silhouette extraction using graph cuts on the GPU. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5716 LNCS 2009. pp. 985-994 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3e95420f3d184c12b7eb4bb678bb74a9,
title = "Real-time online video object silhouette extraction using graph cuts on the GPU",
abstract = "Being able to find the silhouette of an object is a very important front-end processing step for many high-level computer vision techniques, such as Shape-from-Silhouette 3D reconstruction methods, object shape tracking, and pose estimation. Graph cuts have been proposed as a method for finding very accurate silhouettes which can be used as input to such high level techniques, but graph cuts are notoriously computation intensive and slow. Leading CPU implementations can extract a silhouette from a single QVGA image in 100 milliseconds, with performance dramatically decreasing with increased resolution. Recent GPU implementations have been able to achieve performance of 6 milliseconds per image by exploiting the intrinsic properties of the lattice graphs and the hardware model of the GPU. However, these methods are restricted to a subclass of lattice graphs and are not generally applicable. We propose a novel method for graph cuts on the GPU which places no limits on graph configuration and which is able to achieve comparable real-time performance in online video processing scenarios.",
author = "Garrett, {Zachary A.} and Hideo Saito",
year = "2009",
doi = "10.1007/978-3-642-04146-4_105",
language = "English",
isbn = "3642041450",
volume = "5716 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "985--994",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Real-time online video object silhouette extraction using graph cuts on the GPU

AU - Garrett, Zachary A.

AU - Saito, Hideo

PY - 2009

Y1 - 2009

N2 - Being able to find the silhouette of an object is a very important front-end processing step for many high-level computer vision techniques, such as Shape-from-Silhouette 3D reconstruction methods, object shape tracking, and pose estimation. Graph cuts have been proposed as a method for finding very accurate silhouettes which can be used as input to such high level techniques, but graph cuts are notoriously computation intensive and slow. Leading CPU implementations can extract a silhouette from a single QVGA image in 100 milliseconds, with performance dramatically decreasing with increased resolution. Recent GPU implementations have been able to achieve performance of 6 milliseconds per image by exploiting the intrinsic properties of the lattice graphs and the hardware model of the GPU. However, these methods are restricted to a subclass of lattice graphs and are not generally applicable. We propose a novel method for graph cuts on the GPU which places no limits on graph configuration and which is able to achieve comparable real-time performance in online video processing scenarios.

AB - Being able to find the silhouette of an object is a very important front-end processing step for many high-level computer vision techniques, such as Shape-from-Silhouette 3D reconstruction methods, object shape tracking, and pose estimation. Graph cuts have been proposed as a method for finding very accurate silhouettes which can be used as input to such high level techniques, but graph cuts are notoriously computation intensive and slow. Leading CPU implementations can extract a silhouette from a single QVGA image in 100 milliseconds, with performance dramatically decreasing with increased resolution. Recent GPU implementations have been able to achieve performance of 6 milliseconds per image by exploiting the intrinsic properties of the lattice graphs and the hardware model of the GPU. However, these methods are restricted to a subclass of lattice graphs and are not generally applicable. We propose a novel method for graph cuts on the GPU which places no limits on graph configuration and which is able to achieve comparable real-time performance in online video processing scenarios.

UR - http://www.scopus.com/inward/record.url?scp=76249134098&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=76249134098&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-04146-4_105

DO - 10.1007/978-3-642-04146-4_105

M3 - Conference contribution

SN - 3642041450

SN - 9783642041457

VL - 5716 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 985

EP - 994

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -