Bag-of-Event-Models based embeddings for detecting anomalies in surveillance videos
作者：
Highlights：
• Bag-of-Event-Models (BoEM) based embedding is proposed to represent human activities.
• Proposed approach is used for detecting anomalies in surveillance videos.
• Proposed embeddings are of very less dimension.
• Discriminative power & compactness of embeddings lead to improved performance.
• Proposed embeddings are more appropriate for any kind of imbalance sequential data.
摘要
•Bag-of-Event-Models (BoEM) based embedding is proposed to represent human activities.•Proposed approach is used for detecting anomalies in surveillance videos.•Proposed embeddings are of very less dimension.•Discriminative power & compactness of embeddings lead to improved performance.•Proposed embeddings are more appropriate for any kind of imbalance sequential data.
论文关键词：Surveillance videos,Anomaly detection,Bag-of-Event-Models,Motion Boundary Histograms,Hidden Markov Model,Support Vector Machine
论文评审过程：Received 30 November 2020, Revised 5 August 2021, Accepted 27 October 2021, Available online 17 November 2021, Version of Record 23 November 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.116168
