A Machine Learning approach to study soft-tissue facial characteristics as indicators of woman attractiveness

The canons of beauty have undergone substantial changes over the years. Social, cultural and environmental factors influence the perception of beauty. For the face these considerations become even more important, being the key of all social interactions. Several studies have been carried out in this field, to try to give objectivity to these aspects. In this work, starting from a dataset consisting of linear and angular photogrammetric measurements of the faces of 65 women from two different groups, Machine Learning algorithms were trained and tested for the automatic classification of the two groups of individuals. Results were then compared and the predictive power of the adopted classifiers was discussed in terms of sensitivity and specificity.