The effect of orthognathic surgery on facial recognition algorithm analysis

Abstract

The purpose of this study was to evaluate the ability of a publicly available facial recognition application program interface to calculate similarity scores for presurgical and postsurgical photographs of patients who underwent orthognathic surgeries. Our primary objective was to identify which surgical procedure(s) had the greatest effect(s) on the similarity score. Existing photographs for 25 orthodontic-orthognathic patients were analyzed using the application program interface to calculate similarity scores between the presurgical and postsurgical photographs. Photographs from 2 presurgical timepoints were compared as controls. Both relaxed and smiling photographs were included to assess the added impact of a facial pose. Patient characteristics and surgery types were recorded for statistical analysis. Nonparametric Kruskal-Wallis rank-sum tests were performed to analyze the relationship between patient characteristics and similarity scores. Multiple comparisons Wilcoxon rank-sum tests were performed on the statistically significant characteristics. Recognition scores were significantly lower after orthognathic surgery at rest (P=0.009) and smiling (P<0.001). Patients receiving both LeFort I and bilateral sagittal split osteotomy (BSSO) surgeries had a lower median similarity score compared with those that received only BSSO (P=0.009) when comparing relaxed photographs before and after surgery. Similarly, for the score comparing presurgical relaxed photographs to postsurgical smiling photographs, patients that received both surgeries were found to have lower similarity scores compared with those receiving only BSSO (P=0.036). Two-jaw surgeries were associated with a statistically significant decrease in similarity score when compared with 1-jaw procedure. Pose was also found to be a factor influencing similarity scores, especially when comparing presurgical relaxed photographs to postsurgical smiling photographs.