MOD20数据集是由Defence Science and Technology Group创建的，专注于户外多视角人体动作识别。该数据集包含2324个视频片段，总计503086帧，覆盖20种动态人体动作类别。数据来源于YouTube和自主无人机拍摄，视频分辨率为720×720，确保原始比例不变。创建过程涉及从多个视角收集视频，并进行精细的标注工作。MOD20数据集适用于动作识别、监控和态势感知等
This is a pre-processed version of the dataset saved in numpy format. The original dataset is obtained from CMU-MOCAP. The data are 3-dimensional arrays of shape [n_samples, time_steps, n_variables].
