An exploration of new methods for metabolic syndrome examination by infrared thermography and knowledge mining

Metabolic syndrome (MS) is a clinical syndrome with multiple metabolic disorders. As the diagnostic criteria for MS still lacking of imaging laboratory method, this study aimed to explore the differences between healthy people and MS patients through infrared thermography (IRT). However, the observation region of the IRT image is uncertain, and the research tried to solve this problem with the help of knowledge mining technology. 43 MS participants were randomly included through a cross-sectional method, and 43 healthy participants were recruited through number matching. The IRT image of each participant was segmented into the region of interest (ROI) through the preprocessing method proposed in this research, and then the ROI features were granulated by the K-means algorithm to generate the formal background, and finally, the two formal background were separately built into a knowledge graph through the knowledge mining method based on the attribute partial order structure. The baseline data shows that there is no difference in age, gender, and height between the two groups (P > 0.05). The image preprocessing method can segment the IRT image into 18 ROI. Through the K-means method, each group of data can be separately established with a 43 × 36 formal background and generated a knowledge graph. It can be found through knowledge mining and independent-samples T test that the average temperature and maximum temperature difference between the chest and face of the two groups are statistically different (P < 0.01). IRT could reflect the difference between healthy people and MS people. The measurement regions were found by the method of knowledge mining on the premise of unknown. The method proposed in this paper may add a new imaging method for MS laboratory examinations, and at the same time, through knowledge mining, it can also expand a new idea for clinical research of IRT.