An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos

@article{Kiran2018AnOO,
  title={An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos},
  author={Bangalore Ravi Kiran and Dilip Mathew Thomas and Ranjith Parakkal},
  journal={ArXiv},
  year={2018},
  volume={abs/1801.03149}
}
Videos represent the primary source of information for surveillance applications and are available in large amounts but in most cases contain little or no annotation for supervised learning. This article reviews the state-of-the-art deep learning based methods for video anomaly detection and categorizes them based on the type of model and criteria of detection. We also perform simple studies to understand the different approaches and provide the criteria of evaluation for spatio-temporal… 
Improved anomaly detection in surveillance videos based on a deep learning method
  • A. Khaleghi, M. Moin
  • Computer Science
    2018 8th Conference of AI & Robotics and 10th RoboCup Iranopen International Symposium (IRANOPEN)
  • 2018
TLDR
The goal of this article is to propose a new method based on deep learning techniques for anomaly detection in video surveillance cameras, evaluated in the UCSD dataset, and showed an increase in the accuracy of the anomaly detection.
A Critical Study on the Recent Deep Learning Based Semi-Supervised Video Anomaly Detection Methods
TLDR
A novel and deep look at existing methods and results in stating the shortcomings of these approaches are provided, which can be a hint for future works.
Anomaly Detection Based on Latent Feature Training in Surveillance Scenarios
TLDR
It is argued that the constraints in the latent feature space can promote reconstruction; moreover, the optical flow is also considered to introduce temporal information and the results demonstrated the feasibility of the proposed method and the benefit of utilizing information in the hidden feature space.
Self-Supervised Representation Learning for Visual Anomaly Detection
TLDR
This work considers the problem of anomaly detection in images and videos, and presents a new visual anomaly detection technique for videos that identifies the frame indices of a jumbled video sequence allowing it to learn the spatiotemporal features of the video.
A New Semantic and Statistical Distance-Based Anomaly Detection in Crowd Video Surveillance
TLDR
This work investigates a new hybrid visual embedding method based on deep features and a topic model for anomaly detection that demonstrates the effectiveness of this proposed method in anomaly detection.
Anomaly detection in video
TLDR
Novel approaches for learning motion features and modelling normal spatio-temporal dynamics for anomaly detection using deep convolutional neural networks and a sequence-to-sequence encoder-decoder for prediction and reconstruction are presented.
Semi-Supervised Anomaly Detection in Video-Surveillance Scenes in the Wild
TLDR
The proposed approach for anomaly detection in video-surveillance scenes based on a weakly supervised learning algorithm increases the distance between the classification scores of anomalous and normal videos, reducing the number of false negatives.
Anomaly Detection in Videos Using Two-Stream Autoencoder with Post Hoc Interpretability
TLDR
A two-stream approach is introduced that offers an autoencoder-based structure for fast and efficient detection to facilitate anomaly detection from surveillance video without labeled abnormal events and post hoc interpretability of feature map visualization is presented to show the process of feature learning.
Anomaly residual prediction with spatial–temporal and perceptual constraints
TLDR
This work proposes an anomaly detection algorithm based on the state-of-the-art prediction framework, leveraging the gap between frame prediction and its ground truth to detect abnormal events, and develops a new perceptual constraint focusing on high-level information.
Spatio-Temporal Unity Networking for Video Anomaly Detection
TLDR
This study proposes a novel spatio–temporal U-Net for frame prediction using normal events and abnormality detection using prediction error, which combines the benefits of U-Nets in representing spatial information with the capabilities of ConvLSTM for modeling temporal motion data.
...
...

References

SHOWING 1-10 OF 89 REFERENCES
Modeling Representation of Videos for Anomaly Detection using Deep Learning: A Review
TLDR
This paper would like to review the existing methods of modeling video representations using deep learning techniques for the task of anomaly detection and action recognition.
Abnormal Event Detection in Videos using Spatiotemporal Autoencoder
TLDR
This work proposes a spatiotemporal architecture for anomaly detection in videos including crowded scenes that includes two main components, one for spatial feature representation, and one for learning the temporal evolution of the spatial features.
Energy-Based Localized Anomaly Detection in Video Surveillance
TLDR
A unified framework for anomaly detection in video based on the restricted Boltzmann machine, a recent powerful method for unsupervised learning and representation learning, that can detect and localize the abnormalities at pixel level with better accuracy than those of baselines, and achieve competitive performance compared with state-of-the-art approaches.
Spatio-Temporal AutoEncoder for Video Anomaly Detection
TLDR
A novel model called Spatio-Temporal AutoEncoding (ST AutoEncoder or STAE), which utilizes deep neural networks to learn video representation automatically and extracts features from both spatial and temporal dimensions by performing 3-dimensional convolutions, which enhances the motion feature learning in videos.
Spatio-Temporal Anomaly Detection for Industrial Robots through Prediction in Unsupervised Feature Space
TLDR
A new unsupervised learning method to train a deep feature extractor from unlabeled images and shows the use of the learned features in a more traditional classification application for CIFAR-10 dataset.
On the Essence of Unsupervised Detection of Anomalous Motion in Surveillance Videos
TLDR
This paper attempts to fill the knowledge gap by studying the videos tested by existing methods and identifying key components required by an effective unsupervised anomaly detection algorithm, and shows that an unsuper supervised algorithm that captures the key components can be relatively simple and yet perform equally well or better compared to existing methods.
Anomaly detection in crowded scenes
TLDR
A novel framework for anomaly detection in crowded scenes is presented and the proposed representation is shown to outperform various state of the art anomaly detection techniques.
Video anomaly detection using deep incremental slow feature analysis network
TLDR
A deep incremental slow feature analysis (D-IncSFA) network is constructed and applied to directly learning progressively abstract and global high-level representations from raw data sequence, which can precisely detect global anomaly such as crowd panic.
Deep-Cascade: Cascading 3D Deep Neural Networks for Fast Anomaly Detection and Localization in Crowded Scenes
TLDR
It is shown that the proposed novel technique, characterised by a cascade of two cascaded classifiers, performs comparable to current top-performing detection and localization methods on standard benchmarks, but outperforms those in general with respect to required computation time.
...
...