In 1945, during the final months of World War II, a castle in Austria holding an untold number of important works of art burst into flames. Nazis used the castle as a repository for artwork, and historians believe that Nazis intentionally set the fire.
Three of the paintings thought to have been destroyed were the "Faculty Paintings," enormous works by Gustav Klimt, titled "Medicine," "Philosophy," and "Jurisprudence."
Klimt is the Austrian artist best known for his gold-leaf paintings, most notably "The Kiss." All that remains of the "Faculty Paintings" are black and white photographs.
Until now. In partnership with Vienna's Belvedere Museum, Google Arts & Culture used machine learning techniques to recreate what the paintings would have looked like in full color. The project is part of a new online hub dedicated to Klimt that lets viewers explore the artist's life, impact, lesser known works, catalogue in AR, and the Faculty Paintings project.
Researchers took inputs from several sources to create an algorithm that would color the paintings in ways Klimt experts believe could be accurate representations of the originals. One source was newspaper reports that described the colors, such as a vibrant green for the sky in "Philosophy." The other was Klimt's works created around the same time.
“The result for me was surprising because we were able to color it even in the places where we had no knowledge, with machine learning we have good assumptions that Klimt used certain colors,” Dr. Franz Smola, curator at the Belvedere Museum, said in a statement provided by Google.
You can watch how the researchers did it in the video above, and learn more about these works (not so) lost to time, and their creator, here.
Topics Artificial Intelligence
