The Year in Ideas: A to Z
December 12, 2004
In what has become an annual tradition, The New York Times Magazine
takes stock of the passing year by creating a mini-encyclopedia of the
most noteworthy ideas of the previous 12 months. We put out feelers,
fine-tune our journalistic antennae and call on a fleet of reporters
and researchers to scour the infosphere for the most captivating,
baffling, promising and influential ideas from all walks of life --
not just science and technology, politics and policy, but also tattoo
culture and fast-food management, horticulture and shoe design. Once
we separate the wheat of innovation from the chaff of familiar
notions, we offer up the alphabetical harvest now before you: 71 of
the ideas that emerged -- in ways big and small, for better and worse
-- in 2004.
Debunking Photoshop Fakery
Debunking Photoshop Fakery: The sum of these parts creatd an untruthful whole. By RYAN BIGGE
But not for Hany Farid. In May, Farid, a professor of computer science at Dartmouth College, unveiled software that helps determine whether a digital image has been tampered with. Much as art experts detect forgeries by studying the minutia of brush strokes, Farid has devised methods of analyzing the clusters of pixels that make up a digital photo. After crunching the numbers, his program generates a map of the suspect image that calls attention to suspicious areas where tampering may have occurred. Here's how it works. Each pixel in a photo represents a small piece of coded information, and Farid's program looks for patterns of information within the overall composition of the photo. Unaltered images, he discovered, tend to have what you might call naturally occurring patterns of information. Images that have been altered, by contrast, tend to have abnormal patterns of information that, while invisible to the eye, are detectable by computer. Making things easier, most doctored images are produced with common manipulations -- like resizing, duplication, adjusting the contrast and airbrushing.
Farid began thinking about how to authenticate digital imagery after
discovering that photos taken with a digital camera could be
considered admissible evidence in |