ADVERTISMENT
 
 
1 Dec 2008

Virtual You

- 6 Jan 2001
By Jet Propulsion Labratory managed by CalTech   
Page 1 of 2

A digital human-image animation computer system under development at NASA's Jet Propulsion Laboratory, Pasadena, Calif., can use the smallest units of speech, called phonemes, to manipulate a person's facial movements in an image. The system is driven by language rather than by manual animation controls. While development is in the early stages at this point, the eventual result will be photo-realistic animation of a person speaking.

"This is voice-driven, so the image morphs in response to a voice or equivalent input," said principal investigator John Wright of JPL. "Real-time animation is a key step in our development process. Unlike cartoon morphs, this technology uses video and facial movements of real people as its building blocks."

The system, called Digital Personnel, will make it possible to use an image of any human face and make it appear to be speaking naturally. With a videophone, it would be possible to have the option of always portraying the image you wish - no more "bad hair days." or having to wear an oscar dress to impress! Instead, you can appear exactly how you want to appear.A celebrity figure appearing to speak might be another option for an image.

Communication capabilities are being designed for this technology to allow Digital Personnel to work efficiently over telephone as well as data lines. Lower bandwith - the rate of data transmission, or bits per second -- will be used by this system compared to the bandwidth required to transmit real video images. This will allow broader use of the technology while also preserving the appearance of reality in the speaking facial image.

image JPL/CalTech

A person's face being manipulated by Digital Personnel.

"This is a broad-based consumer application," said Jerry Ruddle, G-Tec's vice president for marketing. "The main advantage of this technology is that it produces very realistic animations of people speaking. Because it uses images of actual individuals, the Digital Personnel technology retains the full detail of the face, including subtle shading and shadowing."

The lab "came to us because they saw a good fit," Ruddle said. "Their interest is to see us successfully commercialize the technology," which uses a computer video system that takes dozens of snapshots of a person speaking and breaks the images down to match the sounds that make up whole words.

The technology can then animate a two-dimensional image of any human face so it appears to be speaking naturally. To a consumer encountering a digital person, it would seem more like a face-to-face conversation than an audio-only exchange.

"Digital Personnel is next-generation technology using voice-driven animation of real human images," said Jerry Ruddle, vice president of sales and marketing at Graphco Technologies, Inc., Newtown, Penn. "It will enable us to provide virtual personnel for commercial applications in numerous markets. Web- based customer support, with user-friendly speaking interfaces, is an important application for this technology. Along with other uses for human-like web applications, we project video telephones, broadcasting, distance learning, video games, and motion pictures will also create significant demand for this human-machine interface technology."

 
Have your say
 
can u touch it?
Posted by: guest - 2008-11-04 - 17:23 GMT

Post new comment
Please copy the 5 symbols from this security code image into the box below to submit comment.

I agree to terms and conditions       
 
FirstScience.com

About | Privacy policy | Terms & conditions
© 1995-2008 All rights reserved

Latest News
> Find 1000s more science gadgets & gizmos