A Real-World Dataset for Abnormal Behavior Detection in Turnstile Gates

This dataset contains real-world surveillance data collected from turnstile gate environments for abnormal behavior detection. The dataset includes:- RGB video sequences captured from real operational turnstile gates- Extracted image frames- Various human behaviors, including normal passage and abnormal actions (e.g., tailgating, unauthorized entry, loitering) The data was collected under real-world conditions and is intended for research in:- Computer vision- Video anomaly detection- Human behavior analysis- Smart access control systems This dataset was used in the study:"A TimeSformer-Based Approach with LSTM and Attention for Detecting Abnormal Behaviors in Turnstile Gates"本数据集收录了采集自真实闸机运维环境的监控数据，用于异常行为检测相关研究。 本数据集包含以下内容： - 真实运营场景下闸机采集的RGB视频序列 - 提取得到的图像帧 - 涵盖各类人类行为，包括正常通行与异常动作（例如尾随、未经授权闯入、逗留徘徊） 本数据集采集自真实应用场景，旨在支撑以下研究领域： - 计算机视觉（Computer Vision） - 视频异常检测（Video Anomaly Detection） - 人类行为分析（Human Behavior Analysis） - 智能门禁系统（Smart Access Control Systems） 本数据集曾应用于以下研究：《基于TimeSformer结合LSTM与注意力机制（Attention）的闸机异常行为检测方法》