We are at the dawn of a new era of certainty in art authentication, The Times has suggested in Friday’s leader.
Their optimism stems from the news that Professor Lior Shamir from Lawrence Technological University in Michigan has developed a computer software program that was able to tell, with over 90% accuracy, whether or not a painting was by the abstract expressionist painter Jackson Pollock.
The program, a detailed description of which is soon to be published in the International Journal of Arts and Technology, uses an algorithm that recognises over 4000 types of patterns found in the artist’s work. If there are sufficient similarities in the painting being examined, then the work is likely to be authentic. The technique is not dissimilar to the face recognition technology we wrote about last week.
At the moment the accuracy is not sufficient to be of use to auction houses. However, Professor Shamir has made the software freely available to develop, and in the future such methods for authentication might be welcomed by the industry, which relies on catalogue raisonees, provenance and connoisseurship to solve questions of attribution.
Pollock forgeries are not uncommon, no doubt attracted by the high prices that his work fetch (his auction record is $58, 363, 752). In July last year, a forger was found guilty of selling nearly $1.9m worth of fake Pollocks to art collectors.
The news comes on the back of the recent discovery that Pollock used a structural plan when creating his pictures. In the process of cleaning Alchemy, a 1947 painting in the collection of the Guggenheim, conservators at Opificio delle Pietre Dure of Florence, in collaboration with eleven Italian science institutions, discovered delicate traces of white paint that appear to have formed a grid that Pollock used to guide the composition.
These latest developments confirm the unique nature of Pollock’s work as well as its complexity, and will hopefully act as a deterrent to potential forgers.