Back to 2024 Abstracts
Using Neural Network Facial Recognition Software for Measuring Perceived Age Reduction in Patients Undergoing Blepharoplasty
Jeffrey Khong
*, Alexandra Davis, Oren Wei, Carisa Cooney, Kristen P. Broderick
Department of Plastic and Reconstructive Surgery, Johns Hopkins School of Medicine, Baltimore, MD
Background: Aging-related changes to eyelid morphology such as decreased skin elasticity and ptosis prompt many patients to undergo blepharoplasty surgery. Few studies, however, examine objective measures of surgical success. Therefore, we investigated the ability of a convolutional neural network (CNN) to perceive differences in age before and after blepharoplasty surgery.
Methods: Pre- and post-operative patient blepharoplasty images from inception through December 2023 were extracted from the American Society of Plastic Surgeons website. Patient age, time between photos, gender, and type of procedure were recorded. Patients who underwent concurrent procedures were excluded. FacePlusPlus (www.faceplusplus.com), a CNN, was used to analyze the photos. Two laypersons rated the photos using the Global Aesthetic Improvement Scale (GAIS). Paired t tests compared perceived pre- and post-operative ages, and linear regressions examined factors associated with age reduction.
Results: A total of 107 patients met the inclusion criteria. Ages ranged from 19 to 76 years (mean=52.1 years). On average, the CNN perceived patients preoperatively as younger than their true age (mean=-4.3 years, r=0.747). Blepharoplasty was associated with a postoperative reduction in perceived age (mean=3.2 years, p<0.05). Age reduction was not associated with gender, preoperative age, or procedure type. GAIS scores correlated with perceived age reduction (r=0.33, p<0.05). Patients assessed as older than their true preoperative age had greater perceived age reductions compared to those assessed as younger (6.9 years vs. 1.2 years, p<0.01). The discrepancy between perceived preoperative age and true age correlated with age reduction (r=0.282, p<0.05).
Conclusions: A CNN found that patients were perceived as younger post-blepharoplasty. Perceived age reduction was greatest in patients who appeared older than their true age, particularly for those with larger discrepancies. CNNs may serve as objective tools for assessing cosmetic improvements and may help guide patient preoperative expectations.
Figure 1. Perceived postoperative age reduction of patients appearing younger than their true age (mean=1.2 years, SD=8.4) versus older (mean=6.9 years, SD=8.8) than their true age (p<0.01).
Back to 2024 Abstracts
