Comparative Analysis of Artificial Intelligence-Based Lateral Facial Analysis with Manual and Radiographic Measurements

Objectives The objectives were to generate a baseline dataset using manual and radiographic measurements for lateral facial analysis and perform a comparative assessment with values generated from an Artificial Intelligence (AI) -based smartphone application. Materials and Methods Facial measurements were obtained from 242 individuals using three methods: radiographic, manual and automated measurements generated by the smartphone application. A standardized baseline set was generated by negating inter-observer and intra-observer errors, this was used to analyze the accuracy of measurements generated by the smartphone application with relevant statistical tests. Results On comparison of manual and app values the intra-class correlation coefficients (ICC) was highest for lower facial height (0.97) followed by face ratio (0.93) and upper facial height (0.93). Moderate reliability was found for chin height, upper lip length and nose height, and poor reliability was found for nasal tip and lower lip measurements. Conclusion The AI based smartphone application generates reliable measurements for key facial parameters such as the facial heights and face ratio, supporting its impactful role in Orthodontic diagnosis. Clinical Relevance : AI driven lateral facial analysis can serve as a convenient and cost-effective adjunct for orthodontic evaluation, allowing clinicians to obtain reliable clinical information at the click of a button.