Facial Recognition Neural Networks Confirm Success of Facial Feminization Surgery.

Abstract

BACKGROUND Male-to-female (MtF) transgender patients desire to be identified and treated as female, not only with partners but also in public and social settings. Facial Feminization Surgery (FFS) entails a combination of highly visible changes in facial features which may affect social "first impressions." No study to date has evaluated the impact of FFS on how MtF patients are gender-typed. To study the effectiveness of FFS, we investigated preoperative/postoperative gender-typing using facial recognition neural networks. METHODS In this study, standardized frontal and lateral view preoperative and postoperative images of twenty MtF patients who completed hard and soft tissue FFS procedures were used, along with control images of unoperated cisgender men and women (n=120 images). Four large, public neural networks trained to identify gender based on facial features analyzed the images. Correct gender-typing, improvement in gender-typing (Preop to Postop), and confidence in femininity were analyzed. RESULTS Cisgender Male and Female control frontal images were correctly identified 100% and 98% of the time. Preoperative FFS images, were misgendered 47% of the time (recognized as male) and only correctly identified as female 53% of the time. Postoperative FFS images were gendered correctly 98% of the time; this was an improvement of 45%. Confidence in femininity also improved from a mean Preop FFS of 0.27 to Postop FFS of 0.87. CONCLUSIONS In the first study of its kind, facial recognition neural networks showed improved gender-typing of transgender women from Preop FFS to Postop FFS. This demonstrated the effectiveness of FFS by artificial intelligence methods.