Faces of the Past applies state-of-the-art computer-vision algorithms to one of art history’s oldest questions: who are the people gazing back from unidentified historical portraits?
Using the Netherlands Institute for Art History’s collection of 240,000 digitized artistic portraits, the vast majority which remain anonymous, the project will refine facial-recognition algorithms to identify recurring subjects across painting, print, and photography.
Our results will enrich the institute’s records, support art-historical discovery, and provide open-source tools for cultural heritage research worldwide. More broadly it will enable specialists and the public to discover hidden connections across centuries of Dutch and Flemish art, restoring lost identities, and giving names and stories back to the faces of the past.
