Markovian framework for foreground-background-shadow separation of real world video scenes

Csaba Benedek, T. Szirányi

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

12 Citations (Scopus)

Abstract

In this paper we give a new model for foreground-back-ground-shadow separation. Our method extracts the faithful silhouettes of foreground objects even if they have partly background like colors and shadows are observable on the image. It does not need any a priori information about the shapes of the objects, it assumes only they are not point-wise. The method exploits temporal statistics to characterize the background and shadow, and spatial statistics for the foreground. A Markov Random Field model is used to enhance the accuracy of the separation. We validated our method on outdoor and indoor video sequences captured by the surveillance system of the university campus, and we also tested it on well-known benchmark videos.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages898-907
Number of pages10
Volume3851 LNCS
DOIs
Publication statusPublished - 2006
Event7th Asian Conference on Computer Vision, ACCV 2006 - Hyderabad, India
Duration: Jan 13 2006Jan 16 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3851 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other7th Asian Conference on Computer Vision, ACCV 2006
CountryIndia
CityHyderabad
Period1/13/061/16/06

Fingerprint

Statistics
Spatial Statistics
Benchmarking
Silhouette
Faithful
Color
Surveillance
Random Field
Benchmark
Model
Framework
Background
Object
Universities

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Benedek, C., & Szirányi, T. (2006). Markovian framework for foreground-background-shadow separation of real world video scenes. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3851 LNCS, pp. 898-907). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3851 LNCS). https://doi.org/10.1007/11612032_90

Markovian framework for foreground-background-shadow separation of real world video scenes. / Benedek, Csaba; Szirányi, T.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3851 LNCS 2006. p. 898-907 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3851 LNCS).

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

Benedek, C & Szirányi, T 2006, Markovian framework for foreground-background-shadow separation of real world video scenes. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 3851 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3851 LNCS, pp. 898-907, 7th Asian Conference on Computer Vision, ACCV 2006, Hyderabad, India, 1/13/06. https://doi.org/10.1007/11612032_90
Benedek C, Szirányi T. Markovian framework for foreground-background-shadow separation of real world video scenes. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3851 LNCS. 2006. p. 898-907. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/11612032_90
Benedek, Csaba ; Szirányi, T. / Markovian framework for foreground-background-shadow separation of real world video scenes. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3851 LNCS 2006. pp. 898-907 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{bf6bcccf0de24a76a9bbe8c99089343c,
title = "Markovian framework for foreground-background-shadow separation of real world video scenes",
abstract = "In this paper we give a new model for foreground-back-ground-shadow separation. Our method extracts the faithful silhouettes of foreground objects even if they have partly background like colors and shadows are observable on the image. It does not need any a priori information about the shapes of the objects, it assumes only they are not point-wise. The method exploits temporal statistics to characterize the background and shadow, and spatial statistics for the foreground. A Markov Random Field model is used to enhance the accuracy of the separation. We validated our method on outdoor and indoor video sequences captured by the surveillance system of the university campus, and we also tested it on well-known benchmark videos.",
author = "Csaba Benedek and T. Szir{\'a}nyi",
year = "2006",
doi = "10.1007/11612032_90",
language = "English",
isbn = "3540312196",
volume = "3851 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "898--907",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Markovian framework for foreground-background-shadow separation of real world video scenes

AU - Benedek, Csaba

AU - Szirányi, T.

PY - 2006

Y1 - 2006

N2 - In this paper we give a new model for foreground-back-ground-shadow separation. Our method extracts the faithful silhouettes of foreground objects even if they have partly background like colors and shadows are observable on the image. It does not need any a priori information about the shapes of the objects, it assumes only they are not point-wise. The method exploits temporal statistics to characterize the background and shadow, and spatial statistics for the foreground. A Markov Random Field model is used to enhance the accuracy of the separation. We validated our method on outdoor and indoor video sequences captured by the surveillance system of the university campus, and we also tested it on well-known benchmark videos.

AB - In this paper we give a new model for foreground-back-ground-shadow separation. Our method extracts the faithful silhouettes of foreground objects even if they have partly background like colors and shadows are observable on the image. It does not need any a priori information about the shapes of the objects, it assumes only they are not point-wise. The method exploits temporal statistics to characterize the background and shadow, and spatial statistics for the foreground. A Markov Random Field model is used to enhance the accuracy of the separation. We validated our method on outdoor and indoor video sequences captured by the surveillance system of the university campus, and we also tested it on well-known benchmark videos.

UR - http://www.scopus.com/inward/record.url?scp=33744954909&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33744954909&partnerID=8YFLogxK

U2 - 10.1007/11612032_90

DO - 10.1007/11612032_90

M3 - Conference contribution

AN - SCOPUS:33744954909

SN - 3540312196

SN - 9783540312192

VL - 3851 LNCS

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

SP - 898

EP - 907

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

ER -