Artificially Intelligent Facial Feature Quantification After Facial Filler Injection
Abhishek A. Desai, MD, Ankoor A. Talwar, MBA, Phoebe B. McAuliffe, BS, Martin P. Morris, MBE, Adrienne N. Christopher, MD, Viren Patel, MD, Robyn B. Broach, PhD, Joseph M. Serletti, MD, Ivona Percec, MD.
University of Pennsylvania, Philadelphia, PA, USA.
PURPOSE: Artificial Intelligence (AI) is increasingly ubiquitous in surgery. Commercially available facial AI algorithms can assess aesthetic treatments such as facelift or facial feminization surgery. The purposes of this study are to determine how AI facial-imaging estimates longitudinal age reduction over 90 days following cosmetic filler injection, and to correlate this with patient-reported appraisal of aging.
METHODS: Women aged 40-65 were recruited and injected in a standardized fashion. Patients completed the FACE-QTM quality of life questionnaire and photographed using the VectraŽ M3 3D Imaging device before injection, immediately post-intervention, and at 2-weeks, 4-weeks, and 12-weeks post-injection. Age was estimated using two AI algorithms: Amazon Rekognition and Microsoft Azure Face APIs.
RESULTS: Fillers were administered to sixty-nine women. Both AI algorithms showed no change in age estimate following intervention, or at 14-days, 28-days, or 90-days following intervention (Figure 1). Self-reported appraisal of aging also did not change post-injection, or at 14-days, 28-days, or 90-days post-injection (Figure 1). At 12 weeks, both AI's age estimations were correlated with the patient's perception of aging (ρ=-0.65, ρ=-0.39, p<0.05).
CONCLUSION: Treatment with facial fillers did not change AI age estimation nor patient perception of aging, over 90 days. Over the long-term, there is a negative correlation between AI age estimation and sense of aging. Aging may not be the most discerning metric to study facial filler outcomes, inviting future investigation. 
Back to 2022 Posters