What is Artificial Intelligence?

A Day With Keats, painting by William James Neatby (1860-1910)

A painting masterpiece is an art. But how does a typical computer interprets it? This is quite difficult to do when we talk of high-level concepts of classification, like how, we, humans use our perception abilities and unique taste in art.

But a group of researchers from the University of Girona and the Max Planck Institute in Germany has shown a schematic evidence that a computer can now understand how paintings are done and classified.

They developed mathematical algorithms that provide clues about the artistic style of a painting. How the painter made his brush strokes create a masterpiece that stands to astound many people. The unique composition of colors or certain aesthetic measurements can already be quantified by a computer, but machines are still far from being able to interpret art in the way that people do.

Using low-level classification, the researchers were able to show that their artificial vision algorithms can be programmed to "understand" pictures and compare between artistic strokes.

Lady with flowers by Edmund H. Garrett (1853-1929)

The artificial vision algorithms makes use of low-level pictorial information to determine an artist's brush thickness, the type of material used and the composition of the palette of colors. Although the researchers admit that it's absolutely difficult to copy how humans interpret art, but, according to them "they can look for trends".

Click here to read more about Understanding Art Through Computer, Artificial Vision Algorithms.