Technology, art, and invention were never strange concepts to Steve Perlman. His passion for all three began when he was a boy.

“Even as a toddler, I loved to make things,” Perlman says, noting the excitement he felt when his father taught him how to illuminate a light bulb by connecting it to a battery.

As Perlman grew older, he added photography, movies, and animation to his list of hobbies. In junior high school, he began making short films using clay animation. And a few years later, he built his first computer and designed his own video games for it.

It was in those teenage years that Perlman says he realized his true calling in life: “building technologies to allow people to create artificial worlds and interact with one another.” It’s a goal that has led him to invent something that is certain to take video games and films to new heights. With his Contour Reality Capture system, this IEEE member is changing the face of digital reality as we know it—literally.

The system films actors wearing phosphorescent powder—glow-in-the-dark makeup—under short bursts of fluorescent light. The images captured in the dark and in the light are used to create three-dimensional faces that can be manipulated to make digital characters appear more real than ever before.

“The creative opportunities are enormous,” Perlman says.

The camera system employs a technology called motion-capture, which is used in films and video games. However, current technology captures only the movement, not the appearance of an actor. Contour replicates the entire surface of the face.

“I don’t really think there’s a limit now when it comes to how real digital characters can appear,” Perlman says.

Steve Perlman stands beside his Contour Reality Capture system, whose synchronized digital cameras and flashing fluorescent lights record actors’ expressions to create lifelike faces for animated movies and video games.

Contour captures images at a resolution close to that of high-definition television. Due out in a few months, Contour is being offered through San Francisco-based Mova, a subsidiary of Rearden, a company Perlman founded in 2000, also in San Francisco.

Perlman unveiled Contour in July at the Siggraph computer graphics conference in Boston. It was an immediate hit there, with participants first learning how difficult it is to realistically capture faces, only to find out later “how readily we can capture them,” Perlman says.

THE OLD WAY Today’s motion-capture technology is tedious. To digitize an actor, hundreds of retro-reflective markers are placed in a specific pattern on the actor’s face and body—which takes up to two hours. The actor is then filmed in motion, cameras track the markers, and the 3-D marker position images are uploaded into a computer.

Because the markers track the motion and position of body parts, animators can make the character do whatever they want, realistically, in whatever world they computer-generate. But the marker process has serious limitations.

“It can’t capture the subtleties of human expression,” Perlman says, adding that lips must remain free because the markers interfere with the actor’s speech. Also, it can take animators six to eight weeks to develop a 3-D character because the markers provide a low-resolution representation of the actor. That’s where Contour comes in.

FACIAL FEATURES For Contour, an actor’s face is sponged with a random pattern of phosphorescent makeup, a 10-minute process. The phosphor grains enable the face to be captured in high resolution. In a light-tight room, modified ballasts fire fluorescent lights up to 120 flashes per second—so fast that the room appears steadily lit.

In addition, two arrays of video cameras are set up around the actor’s face. The first array films the actor when the lights are off, recording the glow from the random phosphorescent makeup patterns. The second array of cameras records the actor when the lights are on.

The glowing patterns captured in the dark are correlated to create a precise 3-D surface of the face, while the illuminated images provide texture. With the entire surface of the face captured, Perlman says emotion can be portrayed in a way not currently possible.

“We’re capturing wrinkles, sudden movements of the lips, and the flare of nostrils,” he says, “the subtle things that make human faces expressive.”

The shots then are uploaded into computers that can process the images overnight. Results can be delivered the next morning, ready for editing.

A former engineer at Apple Computer, Stephen G. Perlman has a résumé that includes more than 30 years of technology development. His list of accomplishments includes WebTV, the technology behind QuickTime, and Dish Network’s Dishplayer satellite receiver.

Perlman graduated with a bachelor’s degree in liberal arts in 1983 from Columbia University. Throughout his career, he says, his IEEE membership has been “immeasurably helpful.”

“I do hardware engineering and software, so I spend a lot of time reading IEEE publications,” he says. “The IEEE is a great organization. It has excellent sources of information.”