Multi-Scenario Anomaly Detection (MSAD) dataset
收藏资源简介:
MSAD数据集是由澳大利亚国家大学创建的高分辨率异常检测数据集,包含14个不同场景的监控视频,如道路、商场、公园等。该数据集不仅涵盖了多种正常运动模式,还包括了挑战性的环境变化,如不同的光照和天气条件。创建过程中,数据集从YouTube和实际监控视频中收集,经过精心处理以确保视频质量。MSAD数据集的应用领域广泛,旨在解决监控视频中的异常检测问题,特别是在需要快速适应新场景和视角的情况下。
The MSAD dataset is a high-resolution anomaly detection dataset developed by The Australian National University. It includes surveillance videos across 14 distinct scenarios, such as roads, shopping malls, parks and more. This dataset covers not only diverse normal motion patterns, but also challenging environmental changes including varying lighting and weather conditions. During its development, the dataset was collected from YouTube and real-world surveillance videos, and meticulously processed to ensure video quality. The MSAD dataset has broad application scenarios, aiming to address the anomaly detection task in surveillance videos, especially in scenarios requiring rapid adaptation to new scenes and viewpoints.

- 1Advancing Anomaly Detection: An Adaptation Model and a New Dataset澳大利亚国家大学 · 2024年



