Anomaly Detection in Videos Using Optical Flow and Convolutional Autoencoder

Abstract

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.

One-Class Classification in Images and Videos Using a Convolutional Autoencoder With Compact Embedding

“…Table 2 allows a broad comparison between the state-ofthe-art results and the proposed CAE-CE for video datasets. In the Avenue dataset, our results are better than those achieved by [26], but inferior compared to [14]. However, it is worth mentioning that in [14], different preprocessing stages are applied to the image frames, mainly including optical flow.…”

“…In the Avenue dataset, our results are better than those achieved by [26], but inferior compared to [14]. However, it is worth mentioning that in [14], different preprocessing stages are applied to the image frames, mainly including optical flow. Also, for UCSD Ped2, we achieved satisfactory results, but not as good as those obtained by [12] and [58].…”

“…Notice that the compact representation and dense representation terms are considered synonyms in this work. For the particular case of images and videos, it is known that the features extracted by standard methods may be inefficient when applied directly to OCC problems [11]- [14]. Hand-crafted image descriptors [15]- [19], and [20], such as the Histogram of Oriented Gradients (HOG), may perform well in particular problems where the gradient orientation is relevant for discriminating between the normal and abnormal classes.…”

Video Processing Using Deep Learning Techniques: A Systematic Literature Review

“…This method combines CNN as a feature extractor and LSTM for classification. [105] explain the anomaly detection in realtime videos by using optical-flow convolutional autoencoder and convolutional LSTM. It shows a better performance than the vanilla CNN or DNN based approach for anomaly detection.…”

Violence Detection In Surveillance Videos Using Deep Learning

“…Duman et al [5] propose a system (OF-ConvAE-LSTM) to distinguish anomalies utilizing Convolutional Autoencoder and Convolutional Long Short-Term Memory in an unsupervised way. Other than the deep learning model, the feature extraction stage dependent on thick optical flow is applied in the structure to get the speed and heading data of closer view objects which accomplished accuracy of 89.5 % on avenue dataset ,92.4% on ped1 and 92.9 % on ped2.…”