Deep Learning-Based Evaluation of Gummy Smile: Development and Validation of a Segmentation Model

What this Trial Is About

This research focuses on gummy smile, characterized by excessive gum visibility when smiling. It aims to develop and validate an artificial intelligence (AI) based model to objectively and reliably evaluate gummy smiles from both aesthetic and gum health perspectives. The study uses a unique set of standardized images labeled by an expert to improve clinical accuracy and communication about gummy smiles. Participants will have standardized frontal photos taken with a single device (iPhone 15) under precise conditions to ensure consistent image quality. Smile lines of varying degrees will be labeled by a periodontist using a web-based tool. For high smile lines, gum display will be measured with specialized software and calibrated with a periodontal probe visible in the photos. AI-based analysis results will be compared to manual measurements to assess their accuracy. During the study, participants will have photos taken from a fixed distance while maintaining specific head and eye positions. Researchers will annotate the visible gum areas and measure distances between landmarks on the teeth to evaluate gum visibility. The main outcome is the AI evaluation of gummy smiles over about four months. This study allows healthy volunteers aged 12 and older without missing front upper teeth to participate, with no intervention other than image capture and analysis.

Deep Learning for Gummy Smile Segmentation