The Potential for High-Priority Care Based on Pain Through Facial Expression Detection with Patients Experiencing...

Area of Science:

• Artificial Intelligence
• Medical Technology
• Cardiology

Background:

• Cardiovascular disease (CVD) is a leading cause of mortality, encompassing various heart and vascular disorders.
• Facial expressions can indicate emotional stress, an indirect but significant indicator of CVD.
• Prompt recognition of chest pain is critical to prevent brain cell death during the 'golden hour'.

Purpose of the Study:

• To develop an AI-assisted system for automatic chest pain detection using facial expressions.
• To enhance emergency care services for cardiovascular disease patients.
• To minimize cardiac damage by enabling rapid intervention.

Main Methods:

• Utilized deep learning, specifically the You Only Look Once (YOLO) models, for object detection and recognition.
• Applied a series of YOLO models to detect pain through facial expressions in patients.
• Evaluated YOLOv4, YOLOv6, and YOLOv7 for their efficacy in this task.

Main Results:

• YOLOv4 and YOLOv6 demonstrated superior performance in detecting chest pain via facial expressions compared to YOLOv7.
• Achieved high accuracy rates of 80-100% for YOLOv4 and YOLOv6.
• YOLOv6 exhibited a faster training time than YOLOv4 while maintaining comparable accuracy.

Conclusions:

• Physicians can leverage this AI system to prioritize treatment plans effectively.
• The system aids in reducing the extent of cardiac damage in patients experiencing chest pain.
• Improved patient care and outcomes are expected by optimizing the critical 'golden hour' treatment window.