Unconstrained Facial Beauty Prediction Based on Multi‐scale K ‐Means

Abstract

Facial beauty prediction belongs to an emerging field of human perception nature and rule. Compared with other facial analysis tasks, this task has shown its challenges in pattern recognition and biometric recognition. The algorithm of presented facial beauty prediction requires burden landmark or expensive optimization procedure. We establish a larger database and present a novel method for predicting facial beauty, which is notably superior to previous work in the following aspects: 1) A largescale database with more reasonable distribution has been established and utilized in our experiments; 2) Both female and male facial beauties are analyzed under unconstrained conditions without landmark; 3) Multi-scale apparent features are learned to represent facial beauty which are more expressive and require less computation expenditure. Experimental results demonstrate the accuracy and efficiency of the presented method.