Automatic osteomyelitis area estimation in head CT using anomaly detection

Abstract

In recent years, osteotomy has been used as a treatment for osteomyelitis of the jaw. However, the extent of osteomyelitis that can be determined from preoperative images is ambiguous, which causes problems such as lengthy surgery. Therefore, it is necessary to estimate the resection area with high accuracy. In this study, we proposed a method that combines deep metric learning and anomaly detection to estimate the area of osteomyelitis before surgery. As a result of experiments, we were able to estimate the presence or absence of osteomyelitis with F value of 0.85.