Comparison of Neural Network Model in Deep Learning of Skin Translucency Assessed by Beauty Advisors

A machine learning method that assesses skin translucency, with an accuracy close to that of a beauty advisor, is discussed. A beauty advisor assessed the skin translucency across many facial images captured by a skin analysis system. These facial images were divided into small regions and analyzed using a neural network model. The prediction performance of five well-known neural network models: VGG16, Xception, ResNet50, DenseNet201, and InceptionResNetV2 were compared. The experiments showed that the VGG16 model performed best, achieving a coefficient of determination R2 greater than 0.9. These findings demonstrate the potential of machine learning in addressing the shortage of beauty advisors, further contributing to their training.