Categorizing Pain Intensity Through Facial Recognition Using Deep Learning

Overcrowding in Emergency Departments (ED) is multiplying each year, influencing both the staff and patients and causing a rise in the number of patients in the Intensive Care Unit (ICU). Patients in the ICU require long term observation and extensive monitoring, and with the gap in the ratio of patients to staff, the issue of shortages of staff arises, along with high burnout rates among staff. Though many have tried to undertake this issue, few have succeeded. This paper proposes a deep learning model that aims to reduce the number of ICU and ED patients by categorizing pain intensity levels on a scale of 1–5 in addition to personal use, achieving its goals where no pain achieved a final accuracy of 59.26%, mild pain a final accuracy of 72.73%, moderate pain a final accuracy of 70.34%, severe pain a final accuracy of 64.80%, and very severe pain a final accuracy of 60.69%.