Video anomaly detection with compact feature sets for online performance

Roberto Leyva, Victor Sanchez, Chang Tsun Li

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

Over the past decade, video anomaly detection has been explored with remarkable results. However, research on methodologies suitable for online performance is still very limited. In this paper, we present an online framework for video anomaly detection. The key aspect of our framework is a compact set of highly descriptive features, which is extracted from a novel cell structure that helps to define support regions in a coarse-to-fine fashion. Based on the scene's activity, only a limited number of support regions are processed, thus limiting the size of the feature set. Specifically, we use foreground occupancy and optical flow features. The framework uses an inference mechanism that evaluates the compact feature set via Gaussian Mixture Models, Markov Chains, and Bag-of-Words in order to detect abnormal events. Our framework also considers the joint response of the models in the local spatio-temporal neighborhood to increase detection accuracy. We test our framework on popular existing data sets and on a new data set comprising a wide variety of realistic videos captured by surveillance cameras. This particular data set includes surveillance videos depicting criminal activities, car accidents, and other dangerous situations. Evaluation results show that our framework outperforms other online methods and attains a very competitive detection performance compared with state-of-the-art non-online methods.

Original languageEnglish
Article number7903693
Pages (from-to)3463-3478
Number of pages16
JournalIEEE Transactions on Image Processing
Volume26
Issue number7
DOIs
Publication statusPublished - Jul 2017

Fingerprint

Optical flows
Markov processes
Markov Chains
Accidents
Railroad cars
Cameras
Research Design
Joints
Datasets

Cite this

Leyva, Roberto ; Sanchez, Victor ; Li, Chang Tsun. / Video anomaly detection with compact feature sets for online performance. In: IEEE Transactions on Image Processing. 2017 ; Vol. 26, No. 7. pp. 3463-3478.
@article{d20e2441e8764a73965f646096e49508,
title = "Video anomaly detection with compact feature sets for online performance",
abstract = "Over the past decade, video anomaly detection has been explored with remarkable results. However, research on methodologies suitable for online performance is still very limited. In this paper, we present an online framework for video anomaly detection. The key aspect of our framework is a compact set of highly descriptive features, which is extracted from a novel cell structure that helps to define support regions in a coarse-to-fine fashion. Based on the scene's activity, only a limited number of support regions are processed, thus limiting the size of the feature set. Specifically, we use foreground occupancy and optical flow features. The framework uses an inference mechanism that evaluates the compact feature set via Gaussian Mixture Models, Markov Chains, and Bag-of-Words in order to detect abnormal events. Our framework also considers the joint response of the models in the local spatio-temporal neighborhood to increase detection accuracy. We test our framework on popular existing data sets and on a new data set comprising a wide variety of realistic videos captured by surveillance cameras. This particular data set includes surveillance videos depicting criminal activities, car accidents, and other dangerous situations. Evaluation results show that our framework outperforms other online methods and attains a very competitive detection performance compared with state-of-the-art non-online methods.",
keywords = "Online processing, Video anomaly detection, Video surveillance",
author = "Roberto Leyva and Victor Sanchez and Li, {Chang Tsun}",
note = "Includes bibliographical references.",
year = "2017",
month = "7",
doi = "10.1109/TIP.2017.2695105",
language = "English",
volume = "26",
pages = "3463--3478",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "7",

}

Video anomaly detection with compact feature sets for online performance. / Leyva, Roberto; Sanchez, Victor; Li, Chang Tsun.

In: IEEE Transactions on Image Processing, Vol. 26, No. 7, 7903693, 07.2017, p. 3463-3478.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Video anomaly detection with compact feature sets for online performance

AU - Leyva, Roberto

AU - Sanchez, Victor

AU - Li, Chang Tsun

N1 - Includes bibliographical references.

PY - 2017/7

Y1 - 2017/7

N2 - Over the past decade, video anomaly detection has been explored with remarkable results. However, research on methodologies suitable for online performance is still very limited. In this paper, we present an online framework for video anomaly detection. The key aspect of our framework is a compact set of highly descriptive features, which is extracted from a novel cell structure that helps to define support regions in a coarse-to-fine fashion. Based on the scene's activity, only a limited number of support regions are processed, thus limiting the size of the feature set. Specifically, we use foreground occupancy and optical flow features. The framework uses an inference mechanism that evaluates the compact feature set via Gaussian Mixture Models, Markov Chains, and Bag-of-Words in order to detect abnormal events. Our framework also considers the joint response of the models in the local spatio-temporal neighborhood to increase detection accuracy. We test our framework on popular existing data sets and on a new data set comprising a wide variety of realistic videos captured by surveillance cameras. This particular data set includes surveillance videos depicting criminal activities, car accidents, and other dangerous situations. Evaluation results show that our framework outperforms other online methods and attains a very competitive detection performance compared with state-of-the-art non-online methods.

AB - Over the past decade, video anomaly detection has been explored with remarkable results. However, research on methodologies suitable for online performance is still very limited. In this paper, we present an online framework for video anomaly detection. The key aspect of our framework is a compact set of highly descriptive features, which is extracted from a novel cell structure that helps to define support regions in a coarse-to-fine fashion. Based on the scene's activity, only a limited number of support regions are processed, thus limiting the size of the feature set. Specifically, we use foreground occupancy and optical flow features. The framework uses an inference mechanism that evaluates the compact feature set via Gaussian Mixture Models, Markov Chains, and Bag-of-Words in order to detect abnormal events. Our framework also considers the joint response of the models in the local spatio-temporal neighborhood to increase detection accuracy. We test our framework on popular existing data sets and on a new data set comprising a wide variety of realistic videos captured by surveillance cameras. This particular data set includes surveillance videos depicting criminal activities, car accidents, and other dangerous situations. Evaluation results show that our framework outperforms other online methods and attains a very competitive detection performance compared with state-of-the-art non-online methods.

KW - Online processing

KW - Video anomaly detection

KW - Video surveillance

UR - http://www.scopus.com/inward/record.url?scp=85021710375&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85021710375&partnerID=8YFLogxK

U2 - 10.1109/TIP.2017.2695105

DO - 10.1109/TIP.2017.2695105

M3 - Article

C2 - 28436865

AN - SCOPUS:85021710375

VL - 26

SP - 3463

EP - 3478

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 7

M1 - 7903693

ER -