39 80

Cited 0 times in

A deep neural network-based pain classifier using a photoplethysmography signal

Authors
 Hyunjun Lim  ;  Byeongnam Kim  ;  Gyu-Jeong Noh  ;  Sun K. Yoo 
Citation
 SENSORS, Vol.19(2) : E384, 2019 
Journal Title
 SENSORS 
Issue Date
2019
MeSH
Algorithms ; Deep Learning ; Heart Rate ; Humans ; Middle Aged ; Models, Theoretical ; Neural Networks (Computer) ; Pain/classification ; Pain/physiopathology ; Photoplethysmography ; Postoperative Period ; ROC Curve ; Signal Processing, Computer-Assisted ; Support Vector Machine ; Time Factors
Keywords
bagging ;  deep belief network ;  pain ;  photoplethysmography
Abstract
Side effects occur when excessive or low doses of analgesics are administered compared to the required amount to mediate the pain induced during surgery. It is important to accurately assess the pain level of the patient during surgery. We proposed a pain classifier based on a deep belief network (DBN) using photoplethysmography (PPG). Our DBN learned about a complex nonlinear relationship between extracted PPG features and pain status based on the numeric rating scale (NRS). A bagging ensemble model was used to improve classification performance. The DBN classifier showed better classification results than multilayer perceptron neural network (MLPNN) and support vector machine (SVM) models. In addition, the classification performance was improved when the selective bagging model was applied compared with the use of each single model classifier. The pain classifier based on DBN using a selective bagging model can be helpful in developing a pain classification system.
Files in This Item:
T201902088.pdf Download
DOI
10.3390/s19020384
Appears in Collections:
1. College of Medicine (의과대학) > Dept. of Medical Engineering (의학공학교실) > 1. Journal Papers
Yonsei Authors
Yoo, Sun Kook(유선국) ORCID logo https://orcid.org/0000-0002-6032-4686
URI
https://ir.ymlib.yonsei.ac.kr/handle/22282913/170327
사서에게 알리기
  feedback

qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Browse