Abnormal Event Detection From Videos Using a Two-Stream Recurrent Variational Autoencoder

Yan, Shiyang; Smith, Jeremy S.; Lu, Wenjin; Zhang, Bailing

doi:10.1109/tcds.2018.2883368

Cited by 90 publications

(62 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…UCSDped2 Avenue ShanghaiTech Reconstruction Based Methods CAE [11] 85.0% 80.0% 60.9% SDAE + OCSVM [40] 90.8% --SRNN [25] 92.2% 81.7% 68.0% GAN [32] 93.5% --ConvLSTM-AE [24] 88.1% 77.0% -WTA-CAE + OCSVM [37] 96.6% 82.1% -R-VAE [41] 92.4% 79.6% -PDE-AE [1] 95.4% -72.5% Mem-AE [10] 94.1% 83.3% 71.2% AM-CORR [29] 96.2% 86.9% -AnomalyNet [47] 94 Table 1: AUROC comparison between the proposed VEC and state-of-the-art VAD methods.…”

Section: Methodsmentioning

confidence: 99%

Cloze Test Helps: Effective Video Anomaly Detection via Learning to Complete Video Events

Wang

Cai

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

149

101

View full text Add to dashboard Cite

As a vital topic in media content interpretation, video anomaly detection (VAD) has made fruitful progress via deep neural network (DNN). However, existing methods usually follow a reconstruction or frame prediction routine. They suffer from two gaps: (1) They cannot localize video activities in a both precise and comprehensive manner. (2) They lack sufficient abilities to utilize high-level semantics and temporal context information. Inspired by frequently-used cloze test in language study, we propose a brand-new VAD solution named Video Event Completion (VEC) to bridge gaps above: First, we propose a novel pipeline to achieve both precise and comprehensive enclosure of video activities. Appearance and motion are exploited as mutually complimentary cues to localize regions of interest (RoIs). A normalized spatio-temporal cube (STC) is built from each RoI as a video event, which lays the foundation of VEC and serves as a basic processing unit. Second, we encourage DNN to capture high-level semantics by solving a visual cloze test. To build such a visual cloze test, a certain patch of STC is erased to yield an incomplete event (IE). The DNN learns to restore the original video event from the IE by inferring the missing patch. Third, to incorporate richer motion dynamics, another DNN is trained to infer erased patches' optical flow. Finally, two ensemble strategies using different types of IE and modalities are proposed to boost VAD performance, so as to fully exploit the temporal context and modality information for VAD. VEC can consistently outperform state-ofthe-art methods by a notable margin (typically 1.5%-5% AUROC) on commonly-used VAD benchmarks. Our codes and results can be verified at github.com/ yuguangnudt/ VEC_VAD.

show abstract

Section: Methodsmentioning

confidence: 99%

Cloze Test Helps: Effective Video Anomaly Detection via Learning to Complete Video Events

Wang

Cai

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

149

101

View full text Add to dashboard Cite

show abstract

“…The effectiveness of the proposed framework was tested by comparing our model with six different approaches based on autoencoder. These are Con-vAE [12], ST-AE [2], ConvLSTM-AE [40], Two-Stream R-ConvVAE [41] and WCAE-LSTM [25], and STAN [42]. ConvAE [12] benefits from both fully connected autoencoder with trajectory-based handcrafted spatio-temporal features and convolutional autoencoder.…”

Section: Resultsmentioning

confidence: 99%

Anomaly Detection in Videos Using Optical Flow and Convolutional Autoencoder

Duman

Erdem

2019

IEEE Access

View full text Add to dashboard Cite

Today, public areas, such as airports, hospitals, city centers are monitored by surveillance systems. The widespread use of surveillance systems reduces security concerns while creating an amount of video data that cannot be examined by people in real-time. Therefore, the concept of automatic understanding of video activities has raised the standards of security camera systems. In this paper, we propose a framework (OF-ConvAE-LSTM) to detect anomalies using Convolutional Autoencoder and Convolutional Long Short-Term Memory in an unsupervised manner. Besides the deep learning model, the feature extraction stage based on dense optical flow is applied in the framework to obtain the velocity and direction information of foreground objects. The experiments were carried out on three popular public datasets consisting of Avenue, UCSD Ped1, and UCSD Peds2. The experimental results have shown that the proposed framework models the complex distribution of the pattern of regular motion changes with high accuracy. Besides, this method was observed to outperform state-of-the-art approaches based on unsupervised and semi-supervised deep learning models.INDEX TERMS Abnormal event detection, convolutional autoencoder, long short-term memory, optical flow.

show abstract

“…An example of this is the work carried out by Tran and Hogg where the autoencoder representation is used for detecting anomalies in video [39]. In addition, recurrent autoencoders with LSTM (long short-term memory) layers have also been applied for anomaly detection in video in the work carried out by Yan et al in [40].…”

Section: Related Workmentioning

confidence: 99%

Deep autoencoder for false positive reduction in handgun detection

Vállez

Velasco-Mata

Déniz

2020

Neural Comput & Applic

View full text Add to dashboard Cite

In an object detection system, the main objective during training is to maintain the detection and false positive rates under acceptable levels when the model is run over the test set. However, this typically translates into an unacceptable rate of false alarms when the system is deployed in a real surveillance scenario. To deal with this situation, which often leads to system shutdown, we propose to add a filter step to discard part of the new false positive detections that are typical of the new scenario. This step consists of a deep autoencoder trained with the false alarm detections generated after running the detector over a period of time in the new scenario. Therefore, this step will be in charge of determining whether the detection is a typical false alarm of that scenario or whether it is something anomalous for the autoencoder and, therefore, a true detection. In order to decide whether a detection must be filtered, three different approaches have been tested. The first one uses the autoencoder reconstruction error measured with the mean squared error to make the decision. The other two use the k-NN (k-nearest neighbors) and one-class SVMs (support vector machines) classifiers trained with the autoencoder vector representation. In addition, a synthetic scenario has been generated with Unreal Engine 4 to test the proposed methods in addition to a dataset with real images. The results obtained show a reduction in the number of false positives between 22.5% and 87.2% and an increase in the system’s precision of 1.2%$$-47$$ - 47 % when the autoencoder is applied.

show abstract

Abnormal Event Detection From Videos Using a Two-Stream Recurrent Variational Autoencoder

Cited by 90 publications

References 52 publications

Cloze Test Helps: Effective Video Anomaly Detection via Learning to Complete Video Events

Cloze Test Helps: Effective Video Anomaly Detection via Learning to Complete Video Events

Anomaly Detection in Videos Using Optical Flow and Convolutional Autoencoder

Deep autoencoder for false positive reduction in handgun detection

Contact Info

Product

Resources

About