Human Anomaly Detection
Date
2019
Student ID
Research Supervisor
Editor
Journal Title
Volume
Issue
Journal Title
Journal ISSN
Volume Title
Abstract
Surveillance videos are able to capture a variety of realistic anomalies. In this report, we propose to learn anomalies by exploiting both normal and anomalous videos. To avoid annotating the anomalous segments or clips in training videos, which is very time consuming, we propose to learn anomaly through the deep multiple instance ranking framework by leveraging weakly labeled training videos, i.e. the training labels (anomalous or normal) are at video level instead of clip-level. In our approach, we consider normal and anomalous videos as bags and video segments as instances in multiple instance learning (MIL), and automatically learn a deep anomaly ranking model that predicts high anomaly scores for anomalous video segments. Furthermore, we introduce sparsity and temporal smoothness constraints in the ranking loss function to better localize anomaly during training. We have used real-world surveillance videos, with 13 realistic anomalies such as fighting, road accident, burglary, robbery, etc. as well as normal activities. This dataset can be used for general anomaly detection considering all anomalies in one group and all normal activities in another group. Our experimental results show that our MIL method for anomaly detection achieves significant improvement on anomaly detection performance as compared to the state-of-the-art approaches. The low recognition performance of these baselines reveals that our dataset is very challenging and opens more opportunities for future work.
Description
Keywords
TECHNOLOGY::Electrical engineering, electronics and photonics::Electrical engineering
Citation
Department Name
Electrical and Computer Engineering
Publisher
North South University