2D-gesture

该数据集包含了演员从腰间枪套拔枪、指向目标并返回枪套过程中右手坐标的X和Y值，这些数据是从视频图像中提取的。异常情况被定义为演员未将枪返回枪套的场景。原始数据集仅包含训练集和测试集，其中训练集的30%被用作验证集。时间窗口的长度设置为64，步长为32。任务是对多变量时间序列进行异常检测。

This dataset contains the X and Y coordinate values of the actor's right hand during the sequence of drawing a gun from the waist holster, aiming at a target, and returning the gun to the holster. All data are extracted from video frames. Abnormal cases are defined as scenarios where the actor fails to return the gun to the holster. The original dataset only includes a training set and a test set, where 30% of the training set is utilized as the validation set. The time window length is set to 64, with a stride of 32. The task is anomaly detection for multivariate time series.