We all know that keepin' it real fake has become an art form in itself, but for serious art aficionados, only the real deal will truly suffice. Currently, it takes numerous "art experts" hours to inspect every inkling of a painting to ensure the buyer isn't getting duped, but a new piece of software is looking to realize an idea that's been in the works for years. Igor Berezhnoy at Maastricht University in the Netherlands has created a system which carefully analyzes "uses of color, brush strokes, and canvas type" in order to spot fakes before they fetch millions at auction. By dividing "145 digitized paintings into pixels" and comparing their attributes with nearby pixels, Authentic recognizes uniquities that aren't likely to be expressed in faux editions. In testing, the software performed "as well" as 15 (presumably trained) human volunteers in spotting forgeries, and while the current version works specifically with Van Gogh's works, we're sure a more fleshed out version is on the drawing board.

[Via Slashdot]

Gemstar sues Moxi over guide