Objective Facial Assessment With Artificial Intelligence: Introducing the Facial Aesthetic Index and Facial Youthfulness Index

January 2024 | Volume 23 | Issue 1 | e52 | Copyright © January 2024


Published online December 21, 2023

Sonja Sattler MDa, Konstantin Frank MDb, Martina Kerscher MD PhDc, Sebastian Cotofana MD PhDd, Tatjana Pavicic MDe, Berthold Rzany MD PhDf, Peter Peng MD PhDg, Rainer Pooth MD PhDh

aRosenparkklinik, Darmstadt, Germany 
bDepartment of Hand, Plastic, and Aesthetic Surgery, Ludwig Maximilian University, Munich, Germany 
cDivision of Cosmetic Science, Department of Biochemistry and Molecular Biology, University of Hamburg, Hamburg, Germany
dCentre for Cutaneous Research, Blizard Institute, Queen Mary University of London, London, UK 
ePrivate Practice for Dermatology & Aesthetics, Munich, Germany 
fMedizin am Hauptbahnhof, Wien, Austria 
gDepartment of Dermatology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan 
hClinical Research & Development, ICA Aesthetic Navigation GmbH, Frankfurt, Germany

Abstract

INTRODUCTION

Facial appearance has a substantial impact on psychosocial wellbeing, which may be considered an essential aspect of overall health. Cosmetic rejuvenation using a multimodal approach has been demonstrated to restore a sense of wellness for many patients, with improvements in mental health and self-confidence. Optimal care and treatment outcomes in many fields of medicine rely on objective evaluation metrics rooted in science, and clinicians are increasingly turning to digital health tools to aid in diagnosis and patient management. In aesthetic medicine, assessment is largely subjective and carries a high risk of bias, and dermatology-specific uses account for only a small fraction of available digital tools. A new application using artificial intelligence (AI) has been developed to standardize facial landmarks and characteristics for consistent, unbiased assessment in aesthetic medicine. A detailed analysis of large data sets across gender, age, and ethnicity groups validated by digital images and live assessments contributed to the development of the Facial Aesthetic Index (FAI) and Facial Youthfulness Index (FYI), AI-based algorithms that can identify and prioritize potential interventions for individualized treatment recommendations and provide users with a visible history of treatments and results. As both a diagnostic aid and consultation assistant, the FAI and FYI reflect a holistic impression of facial attractiveness using mathematically selected predictors and have the potential to set a new standard of care in aesthetic rejuvenation.  

Demand for Objective Measurements in Aesthetic Medicine
According to the World Health Organization (WHO), health is not merely the absence of disease or infirmity but is defined as a state of complete physical, mental, and social well-being.1 Facial attractiveness has a demonstrated effect on perceived biological health, mental health, and socioeconomic dimensions.2,3 The abrupt shift to video conferencing during the pandemic has led to increased self-scrutiny and subsequent increase in demand for both surgical and non-surgical aesthetic procedures.4 Research has demonstrated that a multimodal approach to full facial rejuvenation not only improves self-perception of age but also may significantly improve psychological well-being and self-confidence.5 Optimal patient care and satisfaction relies on an objective understanding of aesthetics and beauty, but therein lies the difficulty: there is a lack of objective pre- and post-assessment tools for comprehensive treatment in the largely subjective field of aesthetic medicine.6

Many medical specialties rely on objective diagnostic criteria and outcome assessment based on scientific evidence and measurable treatment response. In aesthetic medicine, there is a lack of standardized, impartial evaluation metrics for assessment and treatment. There are validated scales7,8 that assess the appearance of certain features, such as jowls, nasolabial folds, lateral canthal lines, or infraorbital hollows, but they still require a subjective assignment of severity or grading and fail to provide a complete picture of the aging face. There is no universally accepted definition of beauty and attractiveness.9 Ratios and equations can be used to assess symmetry and proportions for a mathematical appreciation of beauty, but the perception of attractiveness is multidimensional and easily influenced by other factors, such as an individual's self-esteem, apparel, and confidence.10-12 

Clinicians carry inherent biases informed by cultural background, geographic location, familiarity, individual visual environment, social media, peers, and patient population which have an effect on the ability to appreciate "normal" in a patient population and set exact parameters of beauty and attractiveness.9  This may lead to vastly different aesthetic ideals across providers, subjective perceptions about beauty,