Crowd detection with a multiview sampler

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

26 Citations (Scopus)

Abstract

We present a Bayesian approach for simultaneously estimating the number of people in a crowd and their spatial locations by sampling from a posterior distribution over crowd configurations. Although this framework can be naturally extended from single to multiview detection, we show that the naive extension leads to an inefficient sampler that is easily trapped in local modes. We therefore develop a set of novel proposals that leverage multiview geometry to propose global moves that jump more efficiently between modes of the posterior distribution. We also develop a statistical model of crowd configurations that can handle dependencies among people and while not requiring discretization of their spatial locations. We quantitatively evaluate our algorithm on a publicly available benchmark dataset with different crowd densities and environmental conditions, and show that our approach outperforms other state-of-the-art methods for detecting and counting people in crowds.

Original languageEnglish (US)
Title of host publicationComputer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings
PublisherSpringer Verlag
Pages324-337
Number of pages14
EditionPART 5
ISBN (Print)3642155545, 9783642155543
DOIs
StatePublished - Jan 1 2010
Event11th European Conference on Computer Vision, ECCV 2010 - Heraklion, Crete, Greece
Duration: Sep 10 2010Sep 11 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 5
Volume6315 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference11th European Conference on Computer Vision, ECCV 2010
CountryGreece
CityHeraklion, Crete
Period9/10/109/11/10

Fingerprint

Posterior distribution
Configuration
Bayesian Approach
Leverage
Statistical Model
Counting
Jump
Discretization
Benchmark
Sampling
Geometry
Evaluate
Framework
Statistical Models

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ge, W., & Collins, R. T. (2010). Crowd detection with a multiview sampler. In Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings (PART 5 ed., pp. 324-337). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6315 LNCS, No. PART 5). Springer Verlag. https://doi.org/10.1007/978-3-642-15555-0_24
Ge, Weina ; Collins, Robert T. / Crowd detection with a multiview sampler. Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 5. ed. Springer Verlag, 2010. pp. 324-337 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 5).
@inproceedings{b9150be05cc6405e8a7060e3e63936cb,
title = "Crowd detection with a multiview sampler",
abstract = "We present a Bayesian approach for simultaneously estimating the number of people in a crowd and their spatial locations by sampling from a posterior distribution over crowd configurations. Although this framework can be naturally extended from single to multiview detection, we show that the naive extension leads to an inefficient sampler that is easily trapped in local modes. We therefore develop a set of novel proposals that leverage multiview geometry to propose global moves that jump more efficiently between modes of the posterior distribution. We also develop a statistical model of crowd configurations that can handle dependencies among people and while not requiring discretization of their spatial locations. We quantitatively evaluate our algorithm on a publicly available benchmark dataset with different crowd densities and environmental conditions, and show that our approach outperforms other state-of-the-art methods for detecting and counting people in crowds.",
author = "Weina Ge and Collins, {Robert T.}",
year = "2010",
month = "1",
day = "1",
doi = "10.1007/978-3-642-15555-0_24",
language = "English (US)",
isbn = "3642155545",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
number = "PART 5",
pages = "324--337",
booktitle = "Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings",
address = "Germany",
edition = "PART 5",

}

Ge, W & Collins, RT 2010, Crowd detection with a multiview sampler. in Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 5 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 5, vol. 6315 LNCS, Springer Verlag, pp. 324-337, 11th European Conference on Computer Vision, ECCV 2010, Heraklion, Crete, Greece, 9/10/10. https://doi.org/10.1007/978-3-642-15555-0_24

Crowd detection with a multiview sampler. / Ge, Weina; Collins, Robert T.

Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 5. ed. Springer Verlag, 2010. p. 324-337 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6315 LNCS, No. PART 5).

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

TY - GEN

T1 - Crowd detection with a multiview sampler

AU - Ge, Weina

AU - Collins, Robert T.

PY - 2010/1/1

Y1 - 2010/1/1

N2 - We present a Bayesian approach for simultaneously estimating the number of people in a crowd and their spatial locations by sampling from a posterior distribution over crowd configurations. Although this framework can be naturally extended from single to multiview detection, we show that the naive extension leads to an inefficient sampler that is easily trapped in local modes. We therefore develop a set of novel proposals that leverage multiview geometry to propose global moves that jump more efficiently between modes of the posterior distribution. We also develop a statistical model of crowd configurations that can handle dependencies among people and while not requiring discretization of their spatial locations. We quantitatively evaluate our algorithm on a publicly available benchmark dataset with different crowd densities and environmental conditions, and show that our approach outperforms other state-of-the-art methods for detecting and counting people in crowds.

AB - We present a Bayesian approach for simultaneously estimating the number of people in a crowd and their spatial locations by sampling from a posterior distribution over crowd configurations. Although this framework can be naturally extended from single to multiview detection, we show that the naive extension leads to an inefficient sampler that is easily trapped in local modes. We therefore develop a set of novel proposals that leverage multiview geometry to propose global moves that jump more efficiently between modes of the posterior distribution. We also develop a statistical model of crowd configurations that can handle dependencies among people and while not requiring discretization of their spatial locations. We quantitatively evaluate our algorithm on a publicly available benchmark dataset with different crowd densities and environmental conditions, and show that our approach outperforms other state-of-the-art methods for detecting and counting people in crowds.

UR - http://www.scopus.com/inward/record.url?scp=78149352192&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78149352192&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-15555-0_24

DO - 10.1007/978-3-642-15555-0_24

M3 - Conference contribution

AN - SCOPUS:78149352192

SN - 3642155545

SN - 9783642155543

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

SP - 324

EP - 337

BT - Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings

PB - Springer Verlag

ER -

Ge W, Collins RT. Crowd detection with a multiview sampler. In Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 5 ed. Springer Verlag. 2010. p. 324-337. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 5). https://doi.org/10.1007/978-3-642-15555-0_24