Machine learning-powered decoding of facial complexity: a diamond-and-triangle geometry for frontal facial analysis.

Area of Science:

• Aesthetics
• Computer Science
• Anthropology

Background:

• Conventional mathematical laws inadequately represent facial complexity and variability.
• A comprehensive, interracially comparable system is needed for frontal facial analysis.
• Facial attractiveness is influenced by intricate geometric proportions.

Purpose of the Study:

• To develop a machine learning model for analyzing facial attractiveness across different ethnicities.
• To decode esthetic-determining facial traits using a diamond-and-triangle geometric framework.
• To quantify the probability of a face conforming to generally accepted attractiveness standards.

Main Methods:

• Morphological analysis of Caucasian and Asian female faces (attractive vs. ordinary).
• Development of a machine learning model based on a diamond-and-triangle geometric framework.
• Utilizing 21 facial proportion indices derived from the geometry as input for Support Vector Machine (SVM) models.

Main Results:

• SVM models showed excellent efficacy in distinguishing attractive from ordinary faces.
• Attractive faces, across ethnicities, prioritized proportional coordination of facial height with mid-facial width and outer canthal distance.
• Ethnicity-specific variations in the hierarchical importance of other proportion indices were observed.

Conclusions:

• The diamond-and-triangle geometric framework effectively differentiates attractive from ordinary faces, with cross-ethnic similarities.
• The machine learning model provides a valuable tool for cross-ethnic facial analysis.
• This approach offers references for customized treatment planning in facial esthetics.