Spatial-temporal pyramid matching for crowd scene analysis

Hanhe Lin; Jeremiah D. Deng; Brendon J. Woodford

doi:10.1145/2689746.2689751

Spatial-temporal pyramid matching for crowd scene analysis

Hanhe Lin, Jeremiah D. Deng, Brendon J. Woodford

Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Crowd scene analysis has caught significant attention both in academia and industry as it has a great number of potential applications. In this paper, we propose a novel spatial-temporal pyramid matching scheme for crowd scene analysis. Video segments are represented as concatenated histograms of all cells at all pyramid levels with corresponding weights, which reect corresponding matches at finer resolutions are weighted more highly than that found at coarser resolution. Using the classical stochastic gradient descent method, we also propose an online one-class support vector machine algorithm for online anomaly detection scenarios. Extensive experiments have been carried out on two benchmark datasets and the comparison to the state-of-the-art methods validate the advantages of our approach.

Original language	English
Title of host publication	Proceedings of MLSDA 2014
Subtitle of host publication	2nd Workshop on Machine Learning for Sensory Data Analysis
Editors	Ashfaqur Rahman, Jeremiah D. Deng, Jiuyong Li
Place of Publication	New York, NY
Publisher	Association for Computing Machinery (ACM)
Pages	12-18
Number of pages	7
ISBN (Electronic)	978-1-4503-3159-3
DOIs	https://doi.org/10.1145/2689746.2689751
Publication status	Published - 2 Dec 2014
Event	2nd Workshop on Machine Learning for Sensory Data Analysis - Gold Coast, Australia Duration: 2 Dec 2014 → 2 Dec 2014

Conference

Conference	2nd Workshop on Machine Learning for Sensory Data Analysis
Abbreviated title	MLSDA 2014
Country/Territory	Australia
City	Gold Coast
Period	2/12/14 → 2/12/14

Keywords

Anomaly detection
Crowd event recognition
Online one-class support vector machine

ASJC Scopus subject areas

Software
Human-Computer Interaction
Computer Vision and Pattern Recognition
Computer Networks and Communications

Access to Document

10.1145/2689746.2689751

Cite this

@inproceedings{1943201714354572be6928145d2697b5,

title = "Spatial-temporal pyramid matching for crowd scene analysis",

abstract = "Crowd scene analysis has caught significant attention both in academia and industry as it has a great number of potential applications. In this paper, we propose a novel spatial-temporal pyramid matching scheme for crowd scene analysis. Video segments are represented as concatenated histograms of all cells at all pyramid levels with corresponding weights, which reect corresponding matches at finer resolutions are weighted more highly than that found at coarser resolution. Using the classical stochastic gradient descent method, we also propose an online one-class support vector machine algorithm for online anomaly detection scenarios. Extensive experiments have been carried out on two benchmark datasets and the comparison to the state-of-the-art methods validate the advantages of our approach.",

keywords = "Anomaly detection, Crowd event recognition, Online one-class support vector machine",

author = "Hanhe Lin and Deng, {Jeremiah D.} and Woodford, {Brendon J.}",

note = "{\textcopyright} 2014 ACM; 2nd Workshop on Machine Learning for Sensory Data Analysis, MLSDA 2014 ; Conference date: 02-12-2014 Through 02-12-2014",

year = "2014",

month = dec,

day = "2",

doi = "10.1145/2689746.2689751",

language = "English",

pages = "12--18",

editor = "Ashfaqur Rahman and Deng, {Jeremiah D.} and Jiuyong Li",

booktitle = "Proceedings of MLSDA 2014",

publisher = "Association for Computing Machinery (ACM)",

address = "United States",

}

Lin, H, Deng, JD & Woodford, BJ 2014, Spatial-temporal pyramid matching for crowd scene analysis. in A Rahman, JD Deng & J Li (eds), Proceedings of MLSDA 2014: 2nd Workshop on Machine Learning for Sensory Data Analysis. Association for Computing Machinery (ACM), New York, NY, pp. 12-18, 2nd Workshop on Machine Learning for Sensory Data Analysis, Gold Coast, Australia, 2/12/14. https://doi.org/10.1145/2689746.2689751

Spatial-temporal pyramid matching for crowd scene analysis. / Lin, Hanhe; Deng, Jeremiah D.; Woodford, Brendon J.
Proceedings of MLSDA 2014: 2nd Workshop on Machine Learning for Sensory Data Analysis. ed. / Ashfaqur Rahman; Jeremiah D. Deng; Jiuyong Li. New York, NY: Association for Computing Machinery (ACM), 2014. p. 12-18.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Spatial-temporal pyramid matching for crowd scene analysis

AU - Lin, Hanhe

AU - Deng, Jeremiah D.

AU - Woodford, Brendon J.

PY - 2014/12/2

Y1 - 2014/12/2

N2 - Crowd scene analysis has caught significant attention both in academia and industry as it has a great number of potential applications. In this paper, we propose a novel spatial-temporal pyramid matching scheme for crowd scene analysis. Video segments are represented as concatenated histograms of all cells at all pyramid levels with corresponding weights, which reect corresponding matches at finer resolutions are weighted more highly than that found at coarser resolution. Using the classical stochastic gradient descent method, we also propose an online one-class support vector machine algorithm for online anomaly detection scenarios. Extensive experiments have been carried out on two benchmark datasets and the comparison to the state-of-the-art methods validate the advantages of our approach.

AB - Crowd scene analysis has caught significant attention both in academia and industry as it has a great number of potential applications. In this paper, we propose a novel spatial-temporal pyramid matching scheme for crowd scene analysis. Video segments are represented as concatenated histograms of all cells at all pyramid levels with corresponding weights, which reect corresponding matches at finer resolutions are weighted more highly than that found at coarser resolution. Using the classical stochastic gradient descent method, we also propose an online one-class support vector machine algorithm for online anomaly detection scenarios. Extensive experiments have been carried out on two benchmark datasets and the comparison to the state-of-the-art methods validate the advantages of our approach.

KW - Anomaly detection

KW - Crowd event recognition

KW - Online one-class support vector machine

UR - http://www.scopus.com/inward/record.url?scp=84985992245&partnerID=8YFLogxK

U2 - 10.1145/2689746.2689751

DO - 10.1145/2689746.2689751

M3 - Conference contribution

AN - SCOPUS:84985992245

SP - 12

EP - 18

BT - Proceedings of MLSDA 2014

A2 - Rahman, Ashfaqur

A2 - Deng, Jeremiah D.

A2 - Li, Jiuyong

PB - Association for Computing Machinery (ACM)

CY - New York, NY

T2 - 2nd Workshop on Machine Learning for Sensory Data Analysis

Y2 - 2 December 2014 through 2 December 2014

ER -

Spatial-temporal pyramid matching for crowd scene analysis

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this