Two New Methods Lend Scientific Muscle To Detecting Art Forgeries

If you're looking to sell some phony art, your time appears to be running out. Dartmouth researchers can spot art forgeries using neuroscience techniques, while Polish scientists use medical imaging technology to do likewise.

The Dartmouth College team of James Hughes, Daniel Graham, and Daniel Rockmore took an established part of neuroscience - the notion of sparse coding - and applied it to the art world. Sparse coding technology imitates the human visual system by attempting to replicate the ways in which the brain takes complex images and breaks them down into simple patterns thanks to a series of filter neurons. The brain has developed such that it uses relatively few neurons to process common patterns in the natural world, although it needs many more neurons to deal with complex images with which it is unfamiliar.

The team took this technology and applied it to real paintings by Flemish master Peter Bruegel and a set of forgeries:

Essentially, they imagined a visual system that had evolved while being exposed only to Bruegel drawings. Thus, it would process Bruegel drawings using few filters but would have to use many more when looking at anything else-including Bruegel forgeries.

To create this model the researchers obtained a number of genuine and fake Bruegel drawings. They digitally broke the authentic works up into smaller pieces and using sparse coding technology identified the smallest or "sparsest" set of those pieces that could be used as filters. This set of filters essentially quantified Bruegel's unique artistic style by capturing properties repeated throughout the artist's works.

They then processed both the real and fake Bruegel paintings using the system they had created. The real paintings needed far fewer filters in order to process the real ones than it did the fake ones. No matter how good the forgeries were to the naked eye, they broke down into a set of patterns wholly unrecognizable to the sparse coding system that only knew genuine Bruegel works. The research team believes that this represents a technological breakthrough in art authentication in that their method is considerably more accurate than any other yet devised, although they stress their model is not meant to replace the work of art experts.

Meanwhile, at Nicolaus Copernicus University in Torun, Poland, Associate Professor Piotr Targowski has appropriated another trick from medical technology to spot forgeries. This time, the focus is not on outright forgeries, but fake additions to genuine paintings, such as forged artist signatures added years after the fact. Targowski uses Optical Coherence Tomography, originally created to produce 3D images of the layers of the retina, to spot different layers in the painting:

...easel paintings prepared according to traditional techniques consist of multiple layers. The artist, for instance, first applies a glue sizing over the canvas to ensure proper adhesion of later layers. Those layers may include an outline of the painting, the painting itself, layers of semitransparent glazes, and finally transparent varnish.

If the signature or some other inscription is in an out-of-place layer, they can conclude that the element was added at a later date. Targowski and his team have already detected later additions in a couple of 18th and 19th century oil paintings. Perhaps the best news for art lovers is that this method does no physical harm to the paintings. As such, its only their reputations that risk being damaged.

[DUJS Online and OfficialWire]

Send an email to Alasdair Wilkins, the author of this post, at alasdair@io9.com.

publish on Facebook

Upload an image | Add an image URL ×

Image URL: ×

Youtube URL: ×

Choose a file to upload:

×

Tweet this?

Attribute comment to:

Please enter an email address.

Please enter a valid email address.

Dsmvwl | Admin | Promote only | Promote to frontpage | Approve user | Ban user  ×

Loading comments ...

Earlier discussions | Other discussions | Show all discussions | Show featured discussions only | Expand all replies Collapse all replies
Start a new discussion