Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Electrical Engineering

Major Professor

Yasin Yilmaz, Ph.D.

Committee Member

Kwang-Cheng Chen, Ph.D.

Committee Member

Ismail Uysal, Ph.D.

Committee Member

Sudeep Sarkar, Ph.D.

Committee Member

Michael Jones, Ph.D.


Continual Learning, Deep Learning, Interpretability, Video Surveillance, Video Understanding


Anomaly detection in surveillance videos is attracting an increasing amount of attention. Despite the competitive performance of several existing methods, they lack theoretical performance analysis, particularly due to the complex deep neural network architectures used in decision making. Additionally, real-time decision making is an important but mostly neglected factor in this domain. Much of the existing methods that claim to be online, depend on batch or offline processing in practice. Furthermore, several critical tasks such as continual learning, model interpretability and cross-domain adaptability are completely neglected in existing works. Motivated by these research gaps, in this dissertation we discuss our work on real-time video anomaly detection, specifically addressing challenges encountered in a practical implementation. We begin by proposing a multi-objective deep learning module along with a statistical anomaly detection module, and demonstrate its effectiveness on several publicly available data sets. Furthermore, we consider practical challenges such as continual learning and few-shot learning, which humans can easily do but remains to be a significant challenge for machines. A novel algorithm designed for such practical challenges is also proposed. For performance evaluation in this new framework, we introduce a new dataset which is significantly more comprehensive than the existing benchmark datasets, and a new performance metric which takes into account the fundamental temporal aspect of video anomaly detection. Finally, learning from limited data in video surveillance is important for sustainable performance while adapting to new information in a scene over time or adapting to a different scene. In a real-world scene, for an anomaly detection algorithm, all possible nominal patterns and behaviors are not typically available immediately for a single training session. In contrast, labeled nominal data patterns may become available irregularly over a long time horizon, and the anomaly detection algorithm needs to quickly learn such new patterns from limited samples for acceptable performance. Otherwise, it would suffer from frequent false alarms. Cross-domain adaptability (i.e., transfer learning to another surveillance scene) is another task where the anomaly detection algorithm has to quickly learn from limited nominal training data to achieve acceptable performance. Particularly, we study these three problems (few-shot learning, continual learning, cross-domain adaptability) in a multi-task learning setting.