We have proposed a method to detect and quantitatively extract anomalies from surveillance videos. Using our method, anomalies are detected as patterns based on spatio-temporal features that are outliers in new feature space. Conventional anomaly detection methods use features such as tracks or local spatio-temporal features, both of which provide insufficient timing information. Using our method, the principal components of spatio-temporal features of change are extracted from the frames of video sequences of several seconds duration. This enables anomalies based on movement irregularity, both position and speed, to be determined and thus permits the automatic detection of anomal events in sequences of constant length without regard to their start and end. We used a 1-class SVM, which is a non-supervised outlier detection method. The output from the SVM indicates the distance between the outlier and the concentrated base pattern. We demonstrated that the anomalies extracted using our method subjectively matched perceived irregularities in the pattern of movements. Our method is useful in surveillance services because the captured images can be shown in the order of anomality, which significantly reduces the time needed.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Kyoko SUDO, Tatsuya OSAWA, Kaoru WAKABAYASHI, Hideki KOIKE, Kenichi ARAKAWA, "Estimating Anomality of the Video Sequences for Surveillance Using 1-Class SVM" in IEICE TRANSACTIONS on Information,
vol. E91-D, no. 7, pp. 1929-1936, July 2008, doi: 10.1093/ietisy/e91-d.7.1929.
Abstract: We have proposed a method to detect and quantitatively extract anomalies from surveillance videos. Using our method, anomalies are detected as patterns based on spatio-temporal features that are outliers in new feature space. Conventional anomaly detection methods use features such as tracks or local spatio-temporal features, both of which provide insufficient timing information. Using our method, the principal components of spatio-temporal features of change are extracted from the frames of video sequences of several seconds duration. This enables anomalies based on movement irregularity, both position and speed, to be determined and thus permits the automatic detection of anomal events in sequences of constant length without regard to their start and end. We used a 1-class SVM, which is a non-supervised outlier detection method. The output from the SVM indicates the distance between the outlier and the concentrated base pattern. We demonstrated that the anomalies extracted using our method subjectively matched perceived irregularities in the pattern of movements. Our method is useful in surveillance services because the captured images can be shown in the order of anomality, which significantly reduces the time needed.
URL: https://global.ieice.org/en_transactions/information/10.1093/ietisy/e91-d.7.1929/_p
Copy
@ARTICLE{e91-d_7_1929,
author={Kyoko SUDO, Tatsuya OSAWA, Kaoru WAKABAYASHI, Hideki KOIKE, Kenichi ARAKAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Estimating Anomality of the Video Sequences for Surveillance Using 1-Class SVM},
year={2008},
volume={E91-D},
number={7},
pages={1929-1936},
abstract={We have proposed a method to detect and quantitatively extract anomalies from surveillance videos. Using our method, anomalies are detected as patterns based on spatio-temporal features that are outliers in new feature space. Conventional anomaly detection methods use features such as tracks or local spatio-temporal features, both of which provide insufficient timing information. Using our method, the principal components of spatio-temporal features of change are extracted from the frames of video sequences of several seconds duration. This enables anomalies based on movement irregularity, both position and speed, to be determined and thus permits the automatic detection of anomal events in sequences of constant length without regard to their start and end. We used a 1-class SVM, which is a non-supervised outlier detection method. The output from the SVM indicates the distance between the outlier and the concentrated base pattern. We demonstrated that the anomalies extracted using our method subjectively matched perceived irregularities in the pattern of movements. Our method is useful in surveillance services because the captured images can be shown in the order of anomality, which significantly reduces the time needed.},
keywords={},
doi={10.1093/ietisy/e91-d.7.1929},
ISSN={1745-1361},
month={July},}
Copy
TY - JOUR
TI - Estimating Anomality of the Video Sequences for Surveillance Using 1-Class SVM
T2 - IEICE TRANSACTIONS on Information
SP - 1929
EP - 1936
AU - Kyoko SUDO
AU - Tatsuya OSAWA
AU - Kaoru WAKABAYASHI
AU - Hideki KOIKE
AU - Kenichi ARAKAWA
PY - 2008
DO - 10.1093/ietisy/e91-d.7.1929
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E91-D
IS - 7
JA - IEICE TRANSACTIONS on Information
Y1 - July 2008
AB - We have proposed a method to detect and quantitatively extract anomalies from surveillance videos. Using our method, anomalies are detected as patterns based on spatio-temporal features that are outliers in new feature space. Conventional anomaly detection methods use features such as tracks or local spatio-temporal features, both of which provide insufficient timing information. Using our method, the principal components of spatio-temporal features of change are extracted from the frames of video sequences of several seconds duration. This enables anomalies based on movement irregularity, both position and speed, to be determined and thus permits the automatic detection of anomal events in sequences of constant length without regard to their start and end. We used a 1-class SVM, which is a non-supervised outlier detection method. The output from the SVM indicates the distance between the outlier and the concentrated base pattern. We demonstrated that the anomalies extracted using our method subjectively matched perceived irregularities in the pattern of movements. Our method is useful in surveillance services because the captured images can be shown in the order of anomality, which significantly reduces the time needed.
ER -