The following article appeared in The Wall Street Journal Europe.
Cambridge InnoVision software aims to make 3-D models from photos generated
by digital cameras. Museums, take note.
By Brian Truscott, Dow Jones Newswires
Digital cameras seem to be everywhere, what with travelers, proud parents or anyone with a photo to share discovering how easy it is to download, post or send online pictures to anyone around the world.
Martin Weber would like to add one more useful dimension to what a digital camera can do -- or, in his case, a third dimension.
The 31-year-old student, who is about to complete a postgraduate engineering degree at Cambridge University, says he is weeks away from putting the final touches on a prototype software program that will produce three-dimensional models, by simply using photos generated by an off-the-shelf digital camera.
If successful, the ramifications -- and eventual payoff for the student-turned-entrepreneur -- could be substantial, especially if the software is one day sold part-and-parcel with a digital camera.
"If the technology lives up to its promise, the innovation would exploit what is now very much commodity technology," says Neil Macehiter, research director at Ovum, a U.K.-based technology consultancy group. "Digital-camera manufacturers will soon look for ways to tempt consumers to upgrade product, and 3-D photos would grab the attention because it would be very visual."
After all, it's easy to see how travelers would far prefer to have 3-D pictures of London's St. Paul's Cathedral or Michelangelo's statue of David instead of the usual two-dimensional snapshot.
The only requirement? Take four or five pictures of the same object, but from different angles. Plug the digital images into a computer, and a detailed 3-D image will be generated automatically with Mr. Weber's software program. Once the 3-D image is complete, a person will be able to view it from all angles.
The Abyssinian Cat
Mr. Weber's 3-D software program has attracted interest from museums, which would use the camera and software to document and display artifacts such as bones and skulls. The technology "can handle arbitrarily shaped objects, even a model of an ancient Abyssinian cat," says Mike Evans, senior partner at Cambashi Ltd., a Cambridge information-technology consulting firm that has advised Mr. Weber. "It can read dimples, darts and curves."
The electronic gaming and movie industries are also potential clients. Indeed, says Mr. Evans, "it can be used in the entertainment industry, such as for making eye candy for video games."
At the moment, computerized 3-D models are often generated by artists or with fast but expensive laser scanners, which can present a host of problems -- everything from calibration demands to sensor fusion. And there's the hefty price tag: $10,000 (EUR8,700) to $400,000 per laser, plus fees as high as $25,000, depending on the complexity of the work to be done.
A few years ago, Mr. Weber began to see a simpler, low-cost alternative while working on his Ph.D. project, which was all about building computerized three-dimensional models.
"When you look at an issue, read the existing work on the subject, see the problems and then work to solve them, you start to realize how relevant the solution is for a lot of applications," says Mr. Weber, who was raised and educated in Bonn, Germany. "My dream is to use this commercially, not just in the world of academia."
An Ideal Partner
Enter Cambridge Enterprise, which acts as an incubator for students' start-up ventures. A year ago, Mr. Weber started to attend weekly university-sponsored lectures on how to turn an idea into a business venture. He quickly formed a start-up team that included his Malaysian-born wife, Ambili Nair, a Cambridge postgraduate who has since specialized in the biotechnology field. Ms. Nair's business and consultancy experience made her the ideal partner to develop a business plan for her husband's venture, which was incorporated last March under the name Cambridge InnoVision Ltd.
That was the same month Mr. Weber applied for a U.K. patent for his 3-D software program; the lengthy application process should be complete by around next September.
Two other people also joined the team: Nittai Madrid as finance and risk manager, and Rangi Robinson, who will manage the multimedia and sales side of the venture.
Cambridge Enterprise also provided the team with a host of other legal and professional contacts and advisers, while Prof. Robert Cipolla, Mr. Weber's course-work supervisor, and Prof. Andrew Blake, who also works at Microsoft Research on campus, came aboard as technical supervisors.
Mr. Weber and his team plan to move the company forward cautiously, a strategy no doubt influenced by the precipitous rise and nasty fall of many start-up technology companies during the dot-com era in the late 1990s.
"We want to make the prototype work, then show potential clients how it works before we start looking for funding," Ms. Nair says. "We don't want venture capitalists in right at the beginning. Anyway, there isn't that much money around; you have to show that there is an end market for your product first."
In all, Cambridge InnoVision has received some 70 letters of interest from various institutions. Helen Page, the multimedia manager at the National Gallery of Victoria in Melbourne, Australia, wrote to Mr. Weber: "Our photographic services department has been turned around to be completely digital.... This is quite an expensive process. The ability to more easily capture our 3-D objects would be something that we would be extremely interested in."
Charles Lang, a consultant in computer-aided design, helped Mr. Weber write a business plan for a Cambridge entrepreneurship competition. He believes Mr. Weber is right to aim at museums. "For the engineering market, their tech may not be high-enough precision, but it would be OK for museum objects," he says. "For that and some other markets where ultimate precision is not required, like games and such, it could be widely available."
Mr. Weber's prototype, which should be finished by the end of the year, won't be shown to clients until March at the earliest. The reason: Mr. Weber will be immersed in his Ph.D. work from January to March, in order to complete his degree.
Ultimately, if the venture does takes off quickly, that's when the call for additional funding will go out. Clearly, the team isn't keen on giving up control or even a significant stake in the company to external investors until absolutely necessary.
Mr. Weber and his team are aware that myriad companies are trying to perfect computer vision, or to make a PC see objects the way the human brain sees them.
He says the trick is in coming up with the complex mathematical formulas that will ultimately let a computer "model reality" effectively. He equates his software to a computerized "smart" balloon, which automatically shrinks around the object in a digital photo, including interior spaces and holes, until an accurate 3-D image is generated.
"Basically, the software re-forms and changes the model of a computerized object until it agrees with the images you captured on camera," Mr. Weber says.
Reprinted by permission of THE WALL STREET JOURNAL EUROPE, © 2003 Dow Jones & Company, Inc. all rights reserved worldwide
© Cambridge InnoVision Limited