This story is over 5 years old

Computers Are Now Just As Good At Identifying A Painting As The Art Experts

Scientists at Lawrence Technological University have created an algorithm that can determine different styles and periods in art history.

by Kevin Holmes
Oct 4 2012, 4:22pm

When it comes to being an art expect, you usually need to have some kind of background in academia and have spent a lot of time in museums standing in front of paintings, musing over what it all means while pooling your wealth of knowledge about different periods to reflect upon its cultural significance. Or you could invent a computer algorithm and let that do the hard work for you.

Researchers at Lawrence Technological University in Michigan, Lior Shamir and Jane Tarakhovsky, did just that and tested it on 1,000 paintings from 34 famous artists, using identifiers—texture, color, shapes—along with pattern recognition software to deduce similarities and differences between different artworks and painting styles.

It even managed to note different artists from the same movement, linking Baroque painters Vermeer, Rubens, and Rembrandt together, along with artists from other schools like Mannerism, Rococo, Romanticism, Early, High and Northern Renaissance, Impressionism, Post and Neo Impressionism, Abstract Expressionism, Surrealism, and Fauvism—just like any art critic worth their salt would.

So what are the applications? Well, if you’re studying art history, this would certainly be a really great “aid” for helping you write those essays.

[via Gizmag]


Lawrence Technological University