Algorithm distinguishes authentic from fake paintings – An artist who copies a painting, even if the work is his own, will process the colors differently than the original, claim Serbian computer scientists, the technology website Gizmodo reports in a publication.
Scientists have developed an algorithm that is able to distinguish the original painting from the copy. As Milan Rakjović and Miloš Milovanović argue, the technologies and techniques used to analyze an image are now so developed that it is relatively easy, with the right tools, for someone to distinguish the original work, even if the copy was created by the same artist.
In order to prove that the Algorithm they developed is capable of doing what they claim, they hired a Dutch artist to create 7 different works of art and a few days later create their copies.
The scientists then took the 14 paintings in total and examined the shades of red, blue and green using the "wavelet analysis" technique - a process in which the computer gradually examines the blurrier parts of the image.
Based on the results of the analysis, the researchers were able to accurately tell thanks to the Algorithm which paintings were the originals and which were copies.
"For all copies and all paintings, the average global complexity of an original work is greater than the corresponding value of a copy," they said.
The two scientists then used this technique to analyze two nearly identical paintings, which experts believe were created in the 20th century by surrealist artist René Magritte and have puzzled art historians for decades.
The two scientistsclaim to have managed to identify the original painting from the copy, but so far they have not revealed their findings.
Source: skai.gr
