Related Experiment Videos

Automatic Infants' Pain Assessment by Dynamic Facial Representation: Effects of Profile View, Gestational Age,

Ruicong Zhi1,2, Ghada Zamzmi Dmitry Zamzmi3, Dmitry Goldgof4

  • 1School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing 100083, China. zhirc_research@126.com.

Insights

Automated infant pain recognition uses facial expressions, analyzing temporal and geometric features. Gestational age significantly impacts pain assessment accuracy, necessitating individualized models.

Area of Science:

  • Neonatal care and developmental pediatrics
  • Biomedical engineering and machine learning
  • Pain research and clinical assessment

Background:

  • Infants undergoing painful procedures face risks to cognitive, neurological, and brain development.
  • Inability to verbally communicate pain makes objective assessment challenging.
  • Facial expressions are key indicators for automatic pain recognition in non-verbal infants.

Purpose of the Study:

  • To propose a novel dynamic pain facial expression representation and fusion scheme for automatic infant pain assessment.
  • To investigate the influence of individual variables (gestational age, gender, race) and facial views (profile vs. whole face) on pain recognition models.
  • To establish the feasibility of profile-based infant pain assessment.

Main Methods:

  • Developing a fusion scheme combining temporal appearance and geometric facial features for pain expression representation.
  • Constructing and evaluating automatic infant pain assessment models considering gestational age, gender, race, and facial profile.
  • Comparing the performance of profile-based assessment models against whole-face models.

Main Results:

  • The proposed scheme effectively represents dynamic pain facial expressions for automated assessment.
  • Gestational age was identified as the most significant factor influencing pain assessment accuracy.
  • Profile-based infant pain assessment demonstrated feasibility and comparable performance to whole-face assessment.

Conclusions:

  • Automated infant pain recognition is feasible using facial expression analysis, with profile views offering a viable alternative.
  • Gestational age is a critical variable, requiring the development of specific pain assessment models for different gestational ages.
  • This study pioneers the investigation of profile facial views and individual variables in infant pain recognition.

Related Concept Videos