Art criticism and computers
Digital analysis is invading the world of the connoisseur
Jul 30th 2011 | from the print edition
JUDGING artistic styles, and the similarities between them, might be thought one bastion of human skill that machines could never storm. Not so, if Lior Shamir at Lawrence Technological University in Michigan is correct. A paper he has just published in Leonardo suggests that computers may have just as good an eye for style as humans do—and, in some cases, may see connections between artists that human critics have missed.
Dr Shamir, a computer scientist, presented 57 images by each of nine painters—Salvador Dalí, Giorgio de Chirico, Max Ernst, Vasily Kandinsky, Claude Monet, Jackson Pollock, Pierre-Auguste Renoir, Mark Rothko and Vincent van Gogh—to a computer, to see what it made of them. The computer broke the images into a number of so-called numerical descriptors. These descriptors quantified textures and colours, the statistical distribution of edges across a canvas, the distributions of particular types of shape, the intensity of the colour of individual points on a painting, and also the nature of any fractal-like patterns within it (fractals are features that reproduce similar shapes at different scales; the edges of snowflakes, for example).
All told, the computer identified 4,027 different numerical descriptors. Once their values had been established for each of the 513 artworks that had been fed into it, it was ready to do the analysis.
Dr Shamir’s aim was to look for quantifiable ways of distinguishing between the work of different artists. If such things could be established, it might make the task of deciding who painted what a little easier. Such decisions matter because, even excluding deliberate forgeries, there are many paintings in existence that cannot conclusively be attributed to a master rather than his pupils, or that may be honestly made copies whose provenance is now lost.
To look for such distinguishing features, Dr Shamir programmed the computer to use a statistical method that scores the strength of the distance between the values of two or more descriptors for each pair of artists. As a result, he was able to rank each of the 4,027 descriptors by how useful it was at discriminating between artists.
Surprisingly, the values of 19 of the 20 most informative descriptors showed dramatically higher similarities between Van Gogh (left below) and Pollock (right) than between Van Gogh and painters such as Monet and Renoir, who conventional art criticism would think more closely related to Van Gogh’s oeuvre than Pollock’s is. (Dalí and Ernst, by contrast, were farther apart then expected.)
What is interesting, according to Dr Shamir, is that no single feature makes Pollock’s artistic style similar to Van Gogh’s. Instead, the connection is based on a broad set of image-content descriptors which reflect many aspects of the two artists’ styles, including a shared preference for low-level textures and shapes, and similarities in the ways they employed lines and edges.
What was intended, then, as a way of improving the ability to distinguish between different hands has also thrown up a new way of looking for stylistic similarities. Whether Pollock was actually influenced by Van Gogh, or merely happened upon a similar way of doing things through a similar artistic sensibility, is not clear. But it gives art historians a new line of investigation to pursue.