Here's how to detect artistic fakes

Researchers have developed a new technique, known as "sparse coding", to differentiate artistic fakes from real ones.

Researchers have developed a new technique, known as "sparse coding", to differentiate artistic fakes from real ones. It involves building a virtual library of an artist's works and breaking them down into the simplest possible visual elements.

Verifiable works by that artist can be rebuilt using varying proportions of those simple elements, while imitators''works cannot.

Daniel Rockmore of Dartmouth College in the US and his colleagues have developed the method.

They cut each of an artist''s works into 144 pieces, 12 rows and 12 columns. These elements are altered by a computer until some combination of them can recreate each piece from the original artwork. Then the elements are refined until the fewest are required to recreate each piece. Those refined pieces will be unable to reproduce the work of an imitator or a fake.

The experts experimented with the works of Pieter Bruegel the Elder, a 16th century Flemish painter, as part of the project. "Our hope is that it becomes more of what people call technical art history," the BBC News quoted Rockmore, as saying.

He added; "Instead of asking ''was this painting done 40 years after these drawings?', one might instead ask ''how are these statistics evolving over time and what does that say about the working style?'. "For many people those are more central questions, and probably more substantial questions."