ニューラルネットワークを用いた女性の目元・眉元の印象に関する特徴分析

抄録

In recent years, due to the spread of the new coronavirus, it has become common to wear masks in daily life. As a result, makeup around the eyes and eyebrows has become one of the most important aspects of the facial impression one gives to others. The purpose of this study was to clarify the relationship between the proportions of women’s eyes and eyebrows after wearing makeup and their impressions of three items: gender, age, and gorgeousness, using a neural network. First, the proportions were measured from the 2,014 facial images collected, and a neural network was trained using the three impression items of a facial images evaluated by beauty experts as supervised signals. Next, we analyzed the impact of each proportion on the three impression items. Finally, an experiment was conducted with the same experts to verify whether analysis results were adequate. The results revealed how the proportions of the eyes and eyebrows affect the impression one gives to others.