Journal
PAIN
Volume 164, Issue 9, Pages 2029-2035Publisher
LIPPINCOTT WILLIAMS & WILKINS
DOI: 10.1097/j.pain.0000000000002900
Keywords
Pain assessment; Facial expression; Deep learning
Categories
Ask authors/readers for more resources
Current automated pain assessment methods for children are limited to infants and youth, which is impractical in clinical scenarios where children of different ages suffer from postoperative pain. In this study, a large-scale Clinical Pain Expression of Children (CPEC) dataset is presented for assessing postoperative pain in children. A deep learning-based framework called Children Pain Assessment Neural Network (CPANN) is developed to automatically assess postoperative pain based on children's facial expressions. The CPANN achieves an accuracy of 82.1% and a macro-F1 score of 73.9% on the CPEC testing set, demonstrating the effectiveness of the deep learning-based method for automated pain assessment in children.
Current automated pain assessment methods only focus on infants or youth. They are less practical because the children who suffer from postoperative pain in clinical scenarios are in a wider range of ages. In this article, we present a large-scale Clinical Pain Expression of Children (CPEC) dataset for postoperative pain assessment in children. It contains 4104 preoperative videos and 4865 postoperative videos of 4104 children (from 0 to 14 years of age), which are collected from January 2020 to December 2020 in Anhui Provincial Children's Hospital. Moreover, inspired by the dramatic successful applications of deep learning in medical image analysis and emotion recognition, we develop a novel deep learning-based framework to automatically assess postoperative pain according to the facial expression of children, namely Children Pain Assessment Neural Network (CPANN). We train and evaluate the CPANN with the CPEC dataset. The performance of the framework is measured by accuracy and macro-F1 score metrics. The CPANN achieves 82.1% accuracy and 73.9% macro-F1 score on the testing set of CPEC. The CPANN is faster, more convenient, and more objective compared with using pain scales according to the specific type of pain or children's condition. This study demonstrates the effectiveness of deep learning-based method for automated pain assessment in children.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available