The image above looks like an uninspired digital reworking of a beautiful painting, but it's something more. The technique, called wavelet decomposition, shows us the different brushstrokes that make up a painting, and let us see a painter's style in a new way (or an imposter's).

Analysis of style is very complicated, which is why so many art critics occasionally fail to spot a forgery. When looking at a painting, they have to take in an entirely new image and consider the subject, the painter's evolving (or devolving) style and skill, and how the passing years might have affected the artwork. Looking at the painting as a whole can be useful, but now we have the ability to break it down.

A group of mathematicians took a look at famous paintings, as well as their imitators, and decided to focus on the details. They came up with a program that analyzes brushstrokes — particularly their orientation and grouping.

The program takes the angle and orientation of the brushstrokes of a scanned drawing and plots them as a coordinate on a 3D graph. If a subsequent painting has wildly different brushstrokes, it's also plotted on the graph, but will show up far away from the original coordinate. If it has the same brushstrokes, it will be plotted near the coordinate. By looking at which paintings are plotted together, and which are plotted far away from each other, we can guess which are painted by the same person.