UCF-crime 描述数据集 (UCCD)
收藏资源简介:
UCCD是一个专注于视频监控中个体行为描述的数据集,由西北工业大学创建。该数据集包含1012个视频,详细描述了7820个个体的动态行为,涵盖正常与异常场景。每个个体的行为描述包括位置、服装及与其他元素的交互。数据集通过人工标注,确保描述的准确性和客观性。UCCD不仅用于分析监控视频中的个体行为,还适用于动作识别和视频异常检测,为复杂场景下的行为分析提供了丰富的资源。
UCCD is a dataset dedicated to individual behavior description in video surveillance, created by Northwestern Polytechnical University. This dataset comprises 1012 videos, comprehensively documenting the dynamic behaviors of 7820 individuals, covering both normal and abnormal scenarios. The behavior description for each individual includes their location, attire, and interactions with other elements in the scene. The dataset is manually annotated to ensure the accuracy and objectivity of the descriptions. UCCD can be used not only for analyzing individual behaviors in surveillance videos, but also for action recognition and video anomaly detection, providing a rich resource for behavior analysis in complex scenarios.




