From the Department of Physics of Wisdom a new method of analysis that uses algorithms to investigate the properties of the face on which the perception of the beauty of the face depends
We can immediately recognize whether we like a face or not, but what are the properties of the face on which the perception of its beauty depends? The origin and meaning of the beauty of the face have fascinated scholars for centuries, from the ancient Greek canons to the most recent Machine Learning. In fact, there does not seem to be a simple rule, in terms of few features of a face, that can determine perceived beauty with high probability.
A new study by the research team coordinated by Vittorio Loretoof the Department of Physics of Sapienza presented an alternative method of investigation that uses statistical inference techniques for the quantitative study of the phenomenon of perception of facial beauty. The results of the work, published in Scientific Reports , have suggested a subjectivity of the phenomenon of beauty of the face in some ways greater than that proposed so far in the literature. The use of innovative techniques, both experimental and data analysis, allowed the researchers to accurately determine the set of facial changes (made through the digital variants of a portrait) preferred by various subjects.
“Our approach, based on image deformation algorithms and genetic algorithms – explains Vittorio Loreto – allows the experimental subject to ‘sculpt’ his preferred variant of a reference face, navigating through ‘favorite areas’ and converging on specific characteristics inside the so-called face-space or face space “.
The results of the study are compatible with theories on the perception of facial beauty according to which the aesthetic criterion of the various subjects is influenced by their personality and the inference of abstract characteristics is involved in the process of perception of beauty ( personality dimensions) that we unconsciously attribute to others starting from their face. The new method also makes it possible to accurately determine facial features (linear combinations of facial distances) that are significantly more different in faces sculpted by men or women: the latter tend to sculpt variants of a female portrait with a wider face, while men prefer facial changes of the same face with larger eyes, higher cheekbones, lower jaw area and narrower, shorter noses.
The researchers also observed that the differences between male and female subjects (as well as the differences between different subjects) are not evident in the individual facial distances, which are not particularly significant, but in global linear combinations: distances tend to vary together , coherently, from face to face. This reflects the holistic way in which we perceive faces: we do not evaluate the positions of individual facial elements, one at a time, but rather the harmony between the different parts of the face (even the most distant) as a whole. Thus, the way in which the distance between two facial elements is perceived is influenced by the other distances of the face.
Miguel Ibáñez-Berganza concludes first name of the study – proposes an effective experimental analysis scheme in the research area of the perception of the face (which includes the study of identity inference, of age, of psychological attributes and the facial attractivenes. “The perception of the face is the subject of research in increasingly diverse disciplines, such as developmental psychology, evolutionary biology, sociology, neuroscience and machine learning”
Discussion about this post