Detecting Temporal Pain Status of Postoperative Children from Facial Expression

Insufficient analgesia during resuscitation of postoperative children may lead to sequelae such as atelectasis and worsening respiratory function, so timely and accurate detection of pain status in children is essential. At present, the pain status of children is mainly judged by nurses based on facial pain expression scales, which is highly subjective. At the same time, with the increasing shortage of medical resources, untimely detection of pain in children often occurs. In this paper, We built an automatic detection system to detect the pain state of children in real-time, which significantly reduced the workload of medical staff. Specifically, we first design a set of highly flexible facial expression acquisition devices, and then we build a children facial pain expression dataset with the help of clinical experts. After that, we propose an end-to-end children pain detection network, which can automatically evaluate the pain status of children in real-time in an end-to-end framework. Experimental results demonstrate that the evaluation accuracy of the proposed network is higher than that of untrained volunteers and is comparable to the results of clinical experts.