Comparing facial phenotypes using convolutional neural network

Abstract

Automatic analysis of facial phenotypic features from images opens up the possibility of using facial analy- sis to increase the chances that offspring born from gamete donation will resemble their parents. In this paper, experiments in the use of deep learning technologies in the field of facial analysis to advance this goal are presented, including the creation of a database of faces for the evaluation of facial resemblance, the use of facial segmentation, and the detection of three-dimensional phenotypic distances.