Face type recognition method for beauty industry

抄録

Pursuing the perception of beauty is a natural instinct for human beings. Makeup, haircuts, and outfits can make people look more exquisite and attractive. Choosing a suitable fashion style is a form of art, and face types play an important role that is closely related to people's style choices. The importance of human appearance lies in giving people a favorable impression, enhancing positive feelings, and increasing charm. This study aims to meticulously classify people's face types based on their facial features and quickly identify a person's face type, enabling people to understand how to wear makeup and choose outfits to enhance their glamour value. The paper proposes a 16-face types classification method as the formula for achieving beauty. The study employs machine learning (ML), including data acquisition, pre-processing, feature extraction, and classification, to automatically identify face types among sixteen categories and assess accuracy using different samples. To design the supervised machine learning system, the study utilizes the vision transformer (ViT) model for feature extraction and the classical classifier random forest algorithm (RF) for classification. Additionally, the results are compared with other classifiers such as support vector machine (SVM), decision tree (DT), Adaboost, and k-nearest neighbor (k-NN).