心肺复苏姿态评估数据

基于心肺复苏大数据训练平台采集的视频数据，使用ViT Pose算法提取的心肺复苏人体骨骼点数据，包括17个关节在图像空间的绝对坐标值；并使用手工标注方式划分心肺复苏流程关键性动作：{"环境评估","拍肩","判断脉搏","检查呼吸","开放气道","胸外按压","人工呼吸"}；以时间戳为基准，结合提取的骨骼拓扑和手工标注数据，完成对心肺复苏训练过程的关键动作划分和评估。

Video data was collected from a big data training platform for cardiopulmonary resuscitation (CPR). Human skeletal point data of CPR was extracted using the ViT Pose algorithm, which includes the absolute coordinate values of 17 joints in the image space. Key actions in the CPR workflow were manually labeled: Environmental Assessment, Shoulder Tap, Pulse Check, Breathing Check, Airway Opening, Chest Compressions, and Artificial Ventilation. Based on timestamps, combined with the extracted skeletal topology and manually labeled data, key action segmentation and evaluation of the CPR training process were completed.

搜集汇总

数据集介绍

特点

心肺复苏姿态评估数据集基于心肺复苏大数据训练平台采集的视频数据，使用ViT Pose算法提取人体骨骼点数据，并结合手工标注划分心肺复苏流程的关键动作。该数据集主要用于心肺复苏训练考核过程中的关键动作识别和操作标准性评估，数据格式包括excel/csv和json。