Face recognition in historical photo collections

NAISS 2023/22-411


NAISS Small Compute

Principal Investigator:

Anders Hast


Uppsala universitet

Start Date:


End Date:


Primary Classification:

60101: History



This digital humanities project consists of several parts. The main part is to be able to do face recognition among historical photographs, mainly from late 1800 to early 1900. The idea is to later be able to see whether additional photos of a person can help identifying that person as already existing in the database. Historical photos are in many ways different from modern photos. Especially analysis of age and gender fails even more badly for those photos. This is the second part of the project where we would like to see if we can create better models. For both cases we need to do extensive experiments and training new models from scratch and/or using transfer learning. Lastly we also investigate ways of improving photo quality, and thereby also improving the recognition and the analysis of faces.