Abstract
Anomaly detection is one of the most challenging tasks in visual understanding because anomalous events are diverse and complicated. In this paper, we propose a future frame prediction framework and a Multiple Instance Learning (MIL) framework by leveraging attention schemes to learn anomalies. In both frameworks, we utilize the attention-based module to better localize anomalies. Further, we introduce a memory addressing module for the future frame prediction framework, and a novel loss function for the MIL framework, respectively. We also introduce a new multi-view dataset of 170 videos with 10 realistic anomalies that pose a serious threat to security such as intrusion, crowd, accident, weapon, arson, etc., as well as normal activities. The experimental results demonstrate the effectiveness of our methods on the proposed dataset and multiple benchmark datasets. The proposed dataset brings more opportunities and challenges to future work on anomaly detection.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
