Software can identify faces in Holocaust-era photos.
As a reminder, the Holocaust, or “Shoah”, defines the period of the persecution and systematic murder of 6 million Jews from 1933 to 1945. The Nazi State and its collaborators are responsible and organizers.
Find answers to the past
David Patt is a 40 year old software engineer. He now works for Google. Three of his four grandparents are Holocaust survivors in Poland. One of his greatest fears is to come across members of his family in archival photographs of the Holocaust without recognizing them. The engineer’s mother was shot dead during the Nazi invasion, and his grandmother’s brother suffered the same fate when he returned to rescue her. “The rest of the family survived and emigrated to New York after the war”tells the Time of Israel.
This is how he set about creating and developing From Numbers to Names (N2N). It is a facial recognition platform capable of scanning photos and linking them to people living today. The photos concerned are those of pre-war Europe and the Holocaust.
At the base, David Pratt worked alone, now, a team of engineers and scientists accompanies him. So today, in addition to the photos and videos currently available on the platform, the creator has set himself the goal of obtaining access to 700,000 additional photos from the pre-Holocaust era and from the Holocaust.
From Numbers to Names (N2N) is a free non-profit project. At the moment, the software only offers its users the top ten potential matches it finds in its database. It can be descendants of survivors or even celebrities. For example, the United States Holocaust Memorial Museum (USHMM) has already used N2N to search and identify faces in hundreds of thousands of photos.
Note that the results are not 100% reliable but can still help many families. And this, because “the last survivors are dying, and many connections could still be made”insists the creator.
Concretely, all you have to do is click on “select an image” to launch the technology. The user must then choose a file from their computer or phone “containing the cropped photo of a single face”explains David Patt in the Times of Israel interview.
The user must then click on “search” and wait for the software to display 10 photos containing the most similar faces to the one it provided. The engineer points out that “the software works best when searching from photos from around the same time”.
Follow Geeko on Facebook, Youtube and Instagram to not miss any news, tests and tips.