Identification in historical portrait paintings

The knowledge of which person is depicted in an old portrait is not only of interest to art historians, but also to art dealers, collectors and enthusiasts. If that knowledge has been lost with the passage of time, we are generally reliant on expert opinion as to which person was the subject of the painting.

In one of our projects, we successfully used a unique dataset and a unique, innovative method to adapt the principles of modern automated facial recognition systems to the challenge of establishing the identity of persons shown in historical paintings. With the benefit of AI (artificial intelligence), we were able to demonstrate the quality of our method, which also includes a novel assertion of certainty.

With our method, we hope that we can support art historians in their work in the future and to offer new opportunities to the cultural sector (museums, galleries and exhibitions), so that society can see portraits painted centuries ago in a new light.

Overview of our biometrics research