Intelligent Recognition System of Nursing Students’ Procedural Steps of Cardiopulmonary Resuscitation Based on 3D-ResNet

In the traditional training of cardiopulmonary resuscitation (CPR) for nurse students, teachers use observation method to evaluate students' acquisition of operation skills, which not only has high time and manpower cost, but also has strong subjectivity. The automatic recognition of CPR procedural steps is an important prerequisite for intelligent evaluation of CPR skills. Therefore, an automatic recognition system of procedural steps based on 3D-Resnet is proposed for CPR training. First, this study developed an unaware data collection environment for the collection of video data during the CPR training of nurses. Secondly, based on 3D-Resnet, the intelligent recognition system of CPR procedural steps is constructed, and the functional interface of CPR procedural steps analysis report is provided for teachers and students respectively. By constructing an automatic recognition system for CPR procedural steps, the quality and efficiency of CPR training can be significantly improved, which provides a basis for intelligent evaluation in CPR training.