Generative Neural Networks for Anomaly Detection in Crowded Scenes

Tian Wang, Meina Qiao, Zhiwei Lin, Ce Li, Hichem Snoussi, Zhe Liu, Chang Choi

Research output: Contribution to journalArticle

Abstract

Security surveillance is critical to social harmony and people’s peaceful life. It has a great impact on strengthen- ing social stability and life safeguarding. Detecting anoma- ly timely, effectively and efficiently in video surveillance remains challenging. This paper proposes a new approach, called S 2 -VAE, for anomaly detection from video data. The S2-VAE consists of two proposed neural networks: a Stacked Fully Connected Variational AutoEncoder (SF - VAE) and a Skip Convolutional VAE (SC-VAE). The SF -VAE is a shallow generative network to obtain a Gaussian mixture like model to fit the distribution of the actual data. The SC -VAE, as a key component of S 2 - VAE, is a deep generative network to take advantages of CNN, VAE and skip connections. Both SF -VAE and SC-VAE are efficient and effective generative networks and they can achieve better performance for detecting both local abnormal events and global abnormal events. The proposed S 2 -VAE is evaluated using four public datasets. The experimental results show that the S2-VAE outperforms the state-of-the-art algorithms.
LanguageEnglish
Article number1
Pages1-10
Number of pages10
JournalIEEE Transactions on Information Forensics and Security
Volume0
Issue number0
Publication statusAccepted/In press - 13 Oct 2018

Fingerprint

Neural networks

Keywords

  • Spatio-temporal
  • anomaly detection,
  • anomaly detection, Variational AutoEncoder
  • loss function

Cite this

Wang, T., Qiao, M., Lin, Z., Li, C., Snoussi, H., Liu, Z., & Choi, C. (Accepted/In press). Generative Neural Networks for Anomaly Detection in Crowded Scenes. 0(0), 1-10. [1].
Wang, Tian ; Qiao, Meina ; Lin, Zhiwei ; Li, Ce ; Snoussi, Hichem ; Liu, Zhe ; Choi, Chang. / Generative Neural Networks for Anomaly Detection in Crowded Scenes. 2018 ; Vol. 0, No. 0. pp. 1-10.
@article{89b4d1bc769c4d2ca6310a6393e92c87,
title = "Generative Neural Networks for Anomaly Detection in Crowded Scenes",
abstract = "Security surveillance is critical to social harmony and people’s peaceful life. It has a great impact on strengthen- ing social stability and life safeguarding. Detecting anoma- ly timely, effectively and efficiently in video surveillance remains challenging. This paper proposes a new approach, called S 2 -VAE, for anomaly detection from video data. The S2-VAE consists of two proposed neural networks: a Stacked Fully Connected Variational AutoEncoder (SF - VAE) and a Skip Convolutional VAE (SC-VAE). The SF -VAE is a shallow generative network to obtain a Gaussian mixture like model to fit the distribution of the actual data. The SC -VAE, as a key component of S 2 - VAE, is a deep generative network to take advantages of CNN, VAE and skip connections. Both SF -VAE and SC-VAE are efficient and effective generative networks and they can achieve better performance for detecting both local abnormal events and global abnormal events. The proposed S 2 -VAE is evaluated using four public datasets. The experimental results show that the S2-VAE outperforms the state-of-the-art algorithms.",
keywords = "Spatio-temporal, anomaly detection,, anomaly detection, Variational AutoEncoder, loss function",
author = "Tian Wang and Meina Qiao and Zhiwei Lin and Ce Li and Hichem Snoussi and Zhe Liu and Chang Choi",
year = "2018",
month = "10",
day = "13",
language = "English",
volume = "0",
pages = "1--10",
number = "0",

}

Wang, T, Qiao, M, Lin, Z, Li, C, Snoussi, H, Liu, Z & Choi, C 2018, 'Generative Neural Networks for Anomaly Detection in Crowded Scenes', vol. 0, no. 0, 1, pp. 1-10.

Generative Neural Networks for Anomaly Detection in Crowded Scenes. / Wang, Tian; Qiao, Meina; Lin, Zhiwei; Li, Ce; Snoussi, Hichem; Liu, Zhe; Choi, Chang.

Vol. 0, No. 0, 1, 13.10.2018, p. 1-10.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Generative Neural Networks for Anomaly Detection in Crowded Scenes

AU - Wang, Tian

AU - Qiao, Meina

AU - Lin, Zhiwei

AU - Li, Ce

AU - Snoussi, Hichem

AU - Liu, Zhe

AU - Choi, Chang

PY - 2018/10/13

Y1 - 2018/10/13

N2 - Security surveillance is critical to social harmony and people’s peaceful life. It has a great impact on strengthen- ing social stability and life safeguarding. Detecting anoma- ly timely, effectively and efficiently in video surveillance remains challenging. This paper proposes a new approach, called S 2 -VAE, for anomaly detection from video data. The S2-VAE consists of two proposed neural networks: a Stacked Fully Connected Variational AutoEncoder (SF - VAE) and a Skip Convolutional VAE (SC-VAE). The SF -VAE is a shallow generative network to obtain a Gaussian mixture like model to fit the distribution of the actual data. The SC -VAE, as a key component of S 2 - VAE, is a deep generative network to take advantages of CNN, VAE and skip connections. Both SF -VAE and SC-VAE are efficient and effective generative networks and they can achieve better performance for detecting both local abnormal events and global abnormal events. The proposed S 2 -VAE is evaluated using four public datasets. The experimental results show that the S2-VAE outperforms the state-of-the-art algorithms.

AB - Security surveillance is critical to social harmony and people’s peaceful life. It has a great impact on strengthen- ing social stability and life safeguarding. Detecting anoma- ly timely, effectively and efficiently in video surveillance remains challenging. This paper proposes a new approach, called S 2 -VAE, for anomaly detection from video data. The S2-VAE consists of two proposed neural networks: a Stacked Fully Connected Variational AutoEncoder (SF - VAE) and a Skip Convolutional VAE (SC-VAE). The SF -VAE is a shallow generative network to obtain a Gaussian mixture like model to fit the distribution of the actual data. The SC -VAE, as a key component of S 2 - VAE, is a deep generative network to take advantages of CNN, VAE and skip connections. Both SF -VAE and SC-VAE are efficient and effective generative networks and they can achieve better performance for detecting both local abnormal events and global abnormal events. The proposed S 2 -VAE is evaluated using four public datasets. The experimental results show that the S2-VAE outperforms the state-of-the-art algorithms.

KW - Spatio-temporal

KW - anomaly detection,

KW - anomaly detection, Variational AutoEncoder

KW - loss function

M3 - Article

VL - 0

SP - 1

EP - 10

IS - 0

M1 - 1

ER -

Wang T, Qiao M, Lin Z, Li C, Snoussi H, Liu Z et al. Generative Neural Networks for Anomaly Detection in Crowded Scenes. 2018 Oct 13;0(0):1-10. 1.