Context-aware Video Anomaly Detection in Long-Term Datasets
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929311201624064 |
|---|---|
| author | Yang, Zhengye Radke, Richard |
| author_facet | Yang, Zhengye Radke, Richard |
| contents | Video anomaly detection research is generally evaluated on short, isolated benchmark videos only a few minutes long. However, in real-world environments, security cameras observe the same scene for months or years at a time, and the notion of anomalous behavior critically depends on context, such as the time of day, day of week, or schedule of events. Here, we propose a context-aware video anomaly detection algorithm, Trinity, specifically targeted to these scenarios. Trinity is especially well-suited to crowded scenes in which individuals cannot be easily tracked, and anomalies are due to speed, direction, or absence of group motion. Trinity is a contrastive learning framework that aims to learn alignments between context, appearance, and motion, and uses alignment quality to classify videos as normal or anomalous. We evaluate our algorithm on both conventional benchmarks and a public webcam-based dataset we collected that spans more than three months of activity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_07887 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Context-aware Video Anomaly Detection in Long-Term Datasets Yang, Zhengye Radke, Richard Computer Vision and Pattern Recognition Video anomaly detection research is generally evaluated on short, isolated benchmark videos only a few minutes long. However, in real-world environments, security cameras observe the same scene for months or years at a time, and the notion of anomalous behavior critically depends on context, such as the time of day, day of week, or schedule of events. Here, we propose a context-aware video anomaly detection algorithm, Trinity, specifically targeted to these scenarios. Trinity is especially well-suited to crowded scenes in which individuals cannot be easily tracked, and anomalies are due to speed, direction, or absence of group motion. Trinity is a contrastive learning framework that aims to learn alignments between context, appearance, and motion, and uses alignment quality to classify videos as normal or anomalous. We evaluate our algorithm on both conventional benchmarks and a public webcam-based dataset we collected that spans more than three months of activity. |
| title | Context-aware Video Anomaly Detection in Long-Term Datasets |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.07887 |