An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos
This is an incremental overview article for researchers in video surveillance, addressing the lack of annotations in large video datasets.
The paper reviews deep learning methods for unsupervised and semi-supervised anomaly detection in videos, categorizing them by model type and detection criteria, and includes studies to evaluate spatio-temporal approaches.
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 anomaly detection.