clipped from www.oregonlive.comThe man who taught computers to see
Russell Kirsch admits it: Inventing square pixels was a bad idea.
Square dots made pictures fuzzy.
"I started out with a bad idea," Kirsch says, "and that bad idea survived."
Few, however, think of the Portland resident's creation of the pixel itself. Instead, they stand in awe of what those first pixels produced, 50 years ago:
The world's first digital image.
Measuring 176 pixels on each side, the image was a photocopy of Kirsch's infant son, an achievement that Life chronicled in its book "100 Photographs That Changed the World."
That image -- a ghostly, 2-by-2-inch, black-and-white picture of a baby with haunting dark eyes -- spawned a revolution that has yielded one modern miracle after another: CT scanners, satellite imaging, bar-code scanners, desktop publishing, blood-cell analyzers and digital photography, among others.
The revolution was born in a moment of inspiration, when Kirsch asked himself a profound question: "What would happen if a computer could look at the world?"
Now that revolution has reached into the arcana of art. Kirsch and his wife, Joan, an art historian, are using computers to reveal the hidden structures of paintings and the artistic processes of the artists who created them.
Today, Kirsch's original digital image hangs in the Portland Art Museum, 31/2 miles downhill from the comfy West Hills home where Kirsch keeps his office. A copy of the image hangs near his downstairs desk.
clipped from www.oregonlive.comNext to the image is another photo: an older Russell Kirsch with his now-50-year-old son Walden Kirsch. Walden, one of four children, is well-known in the Portland area, having spent 17 years as a general-assignment reporter for KGW-TV.
Walden Kirsch now helps run Intel's worldwide internal news organization out of Hillsboro, where the original photo of himself in his father's arms is wallpaper on his computer screen. (When the photo was scanned into a computer 50 years ago, Russell Kirsch was cropped out.)
"I am delighted that he chose to use for that first photo a real, living, tiny, drooling, pink human being (who happened to be me) instead of, say, a test chart or something," Walden e-mailed this week from Malaysia, where he is traveling on business.
One recent morning, dressed casually in a red shirt, beige suspenders, comfortable slacks and sensible shoes, 77-year-old Russell Kirsch looked more like a mischievous college professor than a world-changing scientist. He spoke with impeccable diction, right down to the subordinate clauses, yet with the joviality of a favorite uncle. He leaned over one of his two Apple MacBooks, calling up Google videos he has made about the early history of computers.
In 1957, Kirsch was a computer programmer -- a job category that, at the time, must have seemed closer to magician than to engineer. Kirsch and an elite team worked with SEAC, the Standards Electronic Automatic Computer, the federal government's first electronic programmable computer.
He had joined the SEAC team in 1951, a year after the National Bureau of Standards finished building the computer. The team used SEAC to solve government problems in fields as diverse as meteorology, navigation and office automation.
But as Kirsch told an oral historian for the National Museum of American History in 1970, he was among a "fortunate happy few" who also had access to the computer for their own private "speculative" experiments, such as artificial intelligence and image processing.
"Sometimes I can confess to having stolen machine time from purportedly more useful products like the thermonuclear weapons calculations and things of this sort," Kirsch told the historian.
In fact, Kirsch said, people with access to computers have a history of using them in unauthorized ways to try interesting experiments that cost-conscious organizations probably wouldn't approve of.
"This, I think, has been a powerful influence on the development of new uses of computers," Kirsch told the historian.
clipped from www.oregonlive.comHe cited his own unauthorized work in acoustic signal generation, which led to work by others in speech synthesis.
Kirsch's work also jump-started the brand-new fields of digital imaging, syntactic pattern recognition and chemical structure searching. He was one of the earliest researchers in artificial intelligence. And he and Joan are among the few experts in the use of computers to study fine arts and ancient petroglyphs.
More than 40 years ago, Kirsch became fascinated with the work linguists were doing to detect "grammars" -- the underlying structures of language.
"I thought, 'I bet you can write grammars for pictures,' " he said.
Joan studied the 140-plus paintings in the "Ocean Park" series by 20th-century American painter Richard Diebenkorn, who was born in Portland. She and Russell, with the help of computers, developed about 42 grammar rules that described the kinds of artistic choices Diebenkorn made in creating his geometrical works.
They generated some "original" Diebenkorns, which were shown to the artist.
"To my amazement," Russell says, "he said he recognized them as some of his compositions."
Diebenkorn phrased it differently in a 1986 interview with Science News: "I looked and felt immediate recognition, and yet it clearly wasn't my work."
Nevertheless, it was the first example of an art historian's ability to document an artistic style with a computer.
Kirsch notes that computers are now being used to create new art, but points out that the past 36,000 years of existing art are still waiting to be discovered anew through computers.
On his MacBook, Kirsch pulls up an image of a sixth-century mosaic made in Ravenna, Italy. The artist had created a sharply delineated picture of a crane, using glass pieces chosen according to color and shape.
Then he pulls up another image of the same crane, this time using the same square pixels he created 50 years ago. The image looks out of focus.
"I did not learn from the smart people in Ravenna 1,500 years ago," he says.
If he had learned, he would have borrowed techniques from the Ravenna mosaicists, instead of using square pixels. And if later computer scientists had learned from Kirsch's mistake, they would not have perpetuated an inferior imaging technique.
"Our desire for the latest thing limits our learning from the past," Kirsch says. "There's so much to be learned from the past, but we're so preoccupied with the future."
Steve Woodward: 503-294-5134; stevewoodward@ news.oregonian.com
No comments:
Post a Comment