Analysis of thermal camera measurements of acute abdominal pain and acute appendicitis patients BY emergency physician

Aim: The aim of this study is to analyze the infrared (thermal imager) images of patients with acute abdominal pain and to diagnose acute appendicitis. Materials and Methods: Patients with acute appendicitis who presented to the emergency department with acute abdominal pain and healthy volunteers without any chronic disease were included in this study. In the patient group, thermal images were taken from the right lower quadrant and left lower quadrant mirror image. The same procedure was repeated for healthy volunteers. As a result of the analysis of these images, the right lower quadrant temperature difference of patients and healthy volunteers was found to be significant (p = 0.01). Then, a model (classifier) was created with the linear support vector machine which is a controlled machine learning method. We tested this model and uploaded the patient and healthy volunteer images that we had not previously taken into the study for testing purposes, and found the accuracy, sensitivity and specificity. Findings: In the patient group, the difference between the right lower quadrant and left lower quadrant temperature was found to be significant (p <0.005). As a result of the analysis of the patient and healthy images we loaded on the linear support machine (artificial memory model) that we trained with these patients, we found the accuracy rate of the artificial memory to be 82.5%, the sensitivity to 96.15%, and the specificity to 57.14%. Results: In this study, it was seen that the artificial memory model created with thermal imager images can help traditional methods in the diagnosis of acute appendicitis. Keywords: Thermal imaging, artificial memory, appendicitis, abdominal pain, emergency medicine.