Morphometric Analysis of Gonial Angle Using ImageJ and Artificial Intelligence: A Cross-sectional Study

Forensic dentistry plays a pivotal role in the identification of individuals by utilizing dental characteristics and records owing to the durability of dental structures, with the mandible being a key element because of its durability and sexual dimorphism. The gonial angle is an important morphometric parameter that varies with age and gender and has been used for identification. However, conventional methods are prone to inter-observer variability, demanding a reproducible approach. The purpose of this study was to compare gonial angle measurements obtained manually using ImageJ software and using an artificial intelligence (AI) model (ChatGPT) on digital orthopantomograms (OPGs) to evaluate sexual dimorphism in gonial angle measurements. A cross-sectional study was conducted using 100 digital OPGs (50 males and 50 females). Gonial angle measurements on both sides were assessed using ImageJ software and AI-based methods. Mean and standard deviation were calculated, and a paired t-test was applied to determine differences between methods. Pearson’s correlation was used to assess the association between age and AI manual discrepancies. AI-derived gonial angles were lower than manual measurements, showing underestimated values on the right side ( p = .0006) and in the mean angle ( p = .0118). Females exhibited higher gonial angles than males, with significant differences between AI and manual measurements in males ( p < .01). No significant correlation was observed between age and AI manual differences ( p > .05). AI-assisted measurements showed comparable trends to manual measurements, although a tendency for underestimation was observed in males. Despite advantages such as accessibility and reduced inter-observer variability, the AI method showed a tendency to underestimate values in males.