This digital humanities research project consists of several parts regarding face recognition. Historical photos are in many ways different from modern photos and the several approaches for this will be investigated such as transfer learning and face 3D frontalisation. Another important task is to create a model for kinship recognition. In the next step we will investigate how deep learning methods can perform dimensionality reduction techniques (t-SNE UMAP etc). We have already concluded that women and men can be separated rather well in 2D space, but we still do not know yet if age, kinship etc can have a similar impact. In any case, we need to do a lot of computing using GPU's since learning is important for all parts of the two research projects I am working in (EB-CRIME and City faces).