Can computers distinguish "creative works of art" with algorithms and rate them?


Computer intelligence is evolving day by day, according to researchers, computers have now developed using visual algorithms to be able to rank historical painters' work based on creative. By saying that,Leonardo da VinciYaMichelangelo,MonetThe contents that the computer evaluated the work such as such as are published to the public.

[1506.00711] Quantifying Creativity in Art Networks

Computer algorithm picks history's 'most creative' paintings (Wired UK)

Professors Ahmed Elgumar and Babak Salehua, computer scientists at Rutgers University, USA, define "creativity" as "product originality and value influence", and using this definition "similar works "How long will there be" based on the basis of "art network" that created what. At this time, the original barometer considered all the elements such as the color and texture of the painting, the subject, and so on.

Then, applying the above measurement method as an algorithm to a database of 600,000 copies of paintings, we compare various paintings from the past to the present. Then, a number of artworks were mapped on the graph as follows. In addition, as it is judged to be "creative" so that it is located at the top of the graph, the more it is judged to be "derivative work", the more points are struck towards the bottom of the graph It has become.

Looking at the graph, Monet's "Shui's Stuffed Straw, Asahi"Edvard Munch"Screaming" and a representative painter of pop artRoy Lichtenstein"Yellow Still Life" along with the top of the graph. This means that the computer regarded it as a "creative work".

On the other hand, however, it is still highly valuedDominique AngleFamous for his works and "thinkers"RodinThe work at the bottom of the graph was "Derivative".

It is impossible to absolutely measure the value of a work of art, so we can not disprove the results derived by the computer. Also note that computers are watching the part called "originality" rather than "essential value" of the work, but "computer algorithms are not innovative," and many art historians are "innovative" and "influence From the fact that the researchers commented that the work which is thought that it is "powerful" is a creative, it is possible that the future that the computer will do is not the critic evaluation of the art is.

in Software,   Art, Posted by darkhorse_log