Software development to identify facial expression of pain in newborn infants

Abstract

Introduction: Although the last years have seen a growing interest in studying me best way of evaluate neonatal pain and stress, it is necessary to develop a better tool to measure or evaluate pain in this age group. Objective: Develop software to identify the newborn infants facial expression of pain by image recognition, to be used in real time, in the intensive care units to help health professionals to monitor its presence and the need of intervention. Method: First a data base of facial images of 30 newborn with gestational age between 28 and 34 weeks will be created, using webcam photographs taken during me blood collection by puncturing the hand's dorsum before, during and after the procedure, with these photos being classified as neutral, pain and distress by 5 neonatologists, blind relative to me moment they were taken. The software will be created in a Windows application development environment oriented to objects in the Delphi's platform. The development of the software will include me import of continuous filming of me newborn, mapping certain facial points from time to time, identifying and e discriminating the facial structures of a 2D image with the newborn's "facial signature". The main points of the image will be standardized while preparing the software: eyes, eyebrows, nose and lips, mapping the newborn's face's geometry and proportions. To allow the software to differentiate the images of pain from those of absence of pain, the comparison between the real time face geometry and that of the image data bank will be programmed, with me analysis of me 8 parameters that determine the presence of pain in the newborn according to the Neonatal Facial Coding System. After a period of taping, a report will be generated showing how many times the patient presented facial features of pain during this interval. (AU)