Ilme – Searching Through Historical Photographs by Face

February 13, 2021

The photographic collections of the National Archives of Estonia are impressive and always worth exploring, with the database FOTIS ( containing over 700.000 digitized and born-digital photographs from the 1840s up to today. It does not only consist of photographic recordings of important public events and figures. The photographs in the archives of for example newspapers, societies or educational institutions may contain photographic depictions of anyone’s predecessors, depending on their life story. Sadly, the information on who is standing before the camera hasn’t always reached the National Archives with the photograph. This is where the new searching tool called Ilme ( comes to the rescue.

Ilme searches not by name but by face. More precisely, it uses the help of artificial intelligence to find the most similar counterparts in FOTIS to the face on an image You upload. This way for example You can try searching for images of Your grandmother. Or take a selfie and look for historical lookalikes of Yourself. Facial recognition technologies are in wide use today, be that by law enforcement agencies or in apps that draw dog ears on Your face. The working principle behind them is generally the same – first the computer searches for where the face is in a picture, then it marks certain points on them like the eyes, nose and mouth. By comparing those to other photographs, the computer then tries to find the closest corresponding faces, hoping that those will be of the same person. Efficient systems do that quite well. Sadly, at this stage, Ilme does not belong to them. The results can be heavily influenced for example by the person’s age in the photo or the angle of their head towards the camera. So one shouldn’t get their hopes too high, but Ilme is still definitely worth a try, for its entertainment as well as practical value.