A double layer background model to detect unusual events

Joaquin Salas, Hugo Jimenez-Hernandez, Jose Joel Gonzalez-Barbosa, Juan B. Hurtado-Ramos, Sandra Canchola

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

2 Citations (Scopus)

Abstract

A double layer background representation to detect novelty in image sequences is shown. The model is capable of handling non-stationary scenarios, such as vehicle intersections. In the first layer, an adaptive pixel appearance background model is computed. Its subtraction with respect to the current image results in a blob description of moving objects. In the second layer, motion direction analysis is performed by a Mixture of Gaussians on the blobs. We have used both layers for representing the usual space of activities and for detecting unusual activity. Our experiments clearly showed that the proposed scheme is able to detect activities such as vehicles running on red light or making forbidden turns.

Original languageEnglish
Title of host publicationAdvanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings
Pages406-416
Number of pages11
StatePublished - 1 Dec 2007
Event9th International Conference on Advanced Concepts for Intelligent Vision Systems, ACIVS 2007 - Delft, Netherlands
Duration: 28 Aug 200731 Aug 2007

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4678 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference9th International Conference on Advanced Concepts for Intelligent Vision Systems, ACIVS 2007
CountryNetherlands
CityDelft
Period28/08/0731/08/07

Fingerprint

Pixels
Red Light
Subtraction
Image Sequence
Moving Objects
Model
Pixel
Intersection
Experiments
Scenarios
Motion
Background
Experiment

Cite this

Salas, J., Jimenez-Hernandez, H., Gonzalez-Barbosa, J. J., Hurtado-Ramos, J. B., & Canchola, S. (2007). A double layer background model to detect unusual events. In Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings (pp. 406-416). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4678 LNCS).
Salas, Joaquin ; Jimenez-Hernandez, Hugo ; Gonzalez-Barbosa, Jose Joel ; Hurtado-Ramos, Juan B. ; Canchola, Sandra. / A double layer background model to detect unusual events. Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings. 2007. pp. 406-416 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3ef345555754472ba14cc6ed78ac63b4,
title = "A double layer background model to detect unusual events",
abstract = "A double layer background representation to detect novelty in image sequences is shown. The model is capable of handling non-stationary scenarios, such as vehicle intersections. In the first layer, an adaptive pixel appearance background model is computed. Its subtraction with respect to the current image results in a blob description of moving objects. In the second layer, motion direction analysis is performed by a Mixture of Gaussians on the blobs. We have used both layers for representing the usual space of activities and for detecting unusual activity. Our experiments clearly showed that the proposed scheme is able to detect activities such as vehicles running on red light or making forbidden turns.",
author = "Joaquin Salas and Hugo Jimenez-Hernandez and Gonzalez-Barbosa, {Jose Joel} and Hurtado-Ramos, {Juan B.} and Sandra Canchola",
year = "2007",
month = "12",
day = "1",
language = "Ingl{\'e}s",
isbn = "9783540746065",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "406--416",
booktitle = "Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings",

}

Salas, J, Jimenez-Hernandez, H, Gonzalez-Barbosa, JJ, Hurtado-Ramos, JB & Canchola, S 2007, A double layer background model to detect unusual events. in Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4678 LNCS, pp. 406-416, 9th International Conference on Advanced Concepts for Intelligent Vision Systems, ACIVS 2007, Delft, Netherlands, 28/08/07.

A double layer background model to detect unusual events. / Salas, Joaquin; Jimenez-Hernandez, Hugo; Gonzalez-Barbosa, Jose Joel; Hurtado-Ramos, Juan B.; Canchola, Sandra.

Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings. 2007. p. 406-416 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4678 LNCS).

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

TY - GEN

T1 - A double layer background model to detect unusual events

AU - Salas, Joaquin

AU - Jimenez-Hernandez, Hugo

AU - Gonzalez-Barbosa, Jose Joel

AU - Hurtado-Ramos, Juan B.

AU - Canchola, Sandra

PY - 2007/12/1

Y1 - 2007/12/1

N2 - A double layer background representation to detect novelty in image sequences is shown. The model is capable of handling non-stationary scenarios, such as vehicle intersections. In the first layer, an adaptive pixel appearance background model is computed. Its subtraction with respect to the current image results in a blob description of moving objects. In the second layer, motion direction analysis is performed by a Mixture of Gaussians on the blobs. We have used both layers for representing the usual space of activities and for detecting unusual activity. Our experiments clearly showed that the proposed scheme is able to detect activities such as vehicles running on red light or making forbidden turns.

AB - A double layer background representation to detect novelty in image sequences is shown. The model is capable of handling non-stationary scenarios, such as vehicle intersections. In the first layer, an adaptive pixel appearance background model is computed. Its subtraction with respect to the current image results in a blob description of moving objects. In the second layer, motion direction analysis is performed by a Mixture of Gaussians on the blobs. We have used both layers for representing the usual space of activities and for detecting unusual activity. Our experiments clearly showed that the proposed scheme is able to detect activities such as vehicles running on red light or making forbidden turns.

UR - http://www.scopus.com/inward/record.url?scp=38149049895&partnerID=8YFLogxK

M3 - Contribución a la conferencia

AN - SCOPUS:38149049895

SN - 9783540746065

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

SP - 406

EP - 416

BT - Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings

ER -

Salas J, Jimenez-Hernandez H, Gonzalez-Barbosa JJ, Hurtado-Ramos JB, Canchola S. A double layer background model to detect unusual events. In Advanced Concepts for Intelligent Vision Systems - 9th International Conference, ACIVS 2007, Proceedings. 2007. p. 406-416. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).