Ten years of Dimensional Imaging

Ten years of Dimensional Imaging
Will Freeman

By Will Freeman

February 4th 2013 at 2:48PM

We shine the spotlight on the facial capture firm after a decade in the industry

Glasgow-based facial capture firm Dimensional Imaging has been around in the industry since its formation at the beginning of 2003, and is now celebrating its tenth anniversary.

The company has offered its services to a number of developers during the past decade, and its facial capture tech has been used in the likes of FIFA and the Dead Island trailer.

To mark the occasion, Develop talked to the firm's CEO Colin Urquhart on what services if offers, how it has managed to keep growing during the past ten years and what its plans are for the future.

Tell us about your company for those who are unfamiliar with your services.
Dimensional Imaging sells high definition facial performance capture systems and software and also offers an on-location facial performance capture service in Europe and North America.

Dimensional Imaging’s approach uses passive stereo photogrammetry to recover actual 3D geometry and colour texture at up to 60 fps. A customer supplied template mesh is then tracked directly through the recovered 3D “scan” sequence ensuring that 3D facial shape and consistent vertex and UV locations are maintained. This approach is able to capture far more of the subtlety of facial performance than alternative marker based or video based approaches.

What has been your biggest accomplishment this year?
One of our biggest accomplishments of the year was to invest in having a 4D capture system permanently based in Los Angeles. This is already starting to pay dividends with several facial performance capture shoots already carried out in Los Angeles and we expect to substantially grow this part of our business in the year ahead.

We have also made our break through into television drama, commercials and music videos which has been tremendously exciting.

What has been your biggest accomplishment to date?
Dimensional Imaging turned ten years old in January 2013, which is a significant accomplishment in itself. In that time, we have grown from a start-up spin out from the Universities of Glasgow and Edinburgh to a highly respected supplier to some of the biggest studios in the business such as Sony Computer Entertainment and Electronic Arts.

Our biggest accomplishment to date has probably been to have our solutions play such a key role in the character modeling pipeline for FIFA.

What projects stand out in your mind as some of the most exciting to be involved in in your time in the industry?
It has been really exciting to see just how extensively our customers such as EA use their DI3D head capture system and DI3D post processing software for such high profile games as EA Sports FIFA. It was also amazingly exciting to play a part in the making of the Dead Island trailer, which, I think it is fair to say, blew everyone involved away by how quickly and how far it spread.

However, the most exciting projects are always the ones we can’t talk about – either because they are not released yet, or because, as has often been the case, the customer insists that we must remain their “secret weapon”. Fortunately after ten years in the business, our secret is staring to leak out!

How has your business serving game developers changed as the industry has changed?
When we first setup the company in January 2003 as a spin out from the Universities of Glasgow and Edinburgh, it was called Virtual Clones and our mission was to “create virtual characters based on the shape and appearance of real life people” using our proprietary passive stereo photogrammetry and re-meshing technologies. We even had a great demo of myself and co-founder Dug Green running around in GTAIII.

However, in retrospect we were ahead of our time and, although we generated some interest from the video game industry, there was no pressing need for our solutions then because of the relative simplicity of the graphics ten years ago. So we changed the name of the company to Dimensional Imaging and concentrated on serving our research customers for the next few years.

But eventually, video game graphics progressed to the stage where it was becoming increasingly costly to produce the required level of realism from traditional methods such as hand modeling heads from reference photos, especially for franchised games like sports sims, where it is important to have good likenesses of famous people in the game.

We made break-through sales of “DI3D” face and head capture systems in 2008-to-2009 to several large game studios and haven’t looked back since.


 
Our “DI4D” facial performance capture system has actually followed a similar path to our 3D solutions. DI4D was probably ahead of its time when we first showed it at Siggraph in San Diego in 2007.

But since then, the level of realism required in facial animation has increased enormously, to the point where studios are now finding it increasingly hard to get the level of realism they require from traditional methods of facial motion capture.

We had our breakthrough for 4D capture in 2011, when Axis Animation asked us to capture the facial performances for the now famous Dead Island trailer.

Since then, interest in our facial performance capture solutions from game studios has also taken off. We now have several large game studios using their own DI4D capture systems for internal projects. We also now offer an on-location 4D facial performance capture service and have recently invested in a second system based in Los Angeles to meet increasing demand there.
 
On a similar note, how has the role and status of facial animation changed in your time in the industry?
When we started in the industry, facial animation was a relatively minor and unimportant aspect of games and game development.

By contrast, the realism of facial animation now expected demands a large proportion of game assets and a large slice of development budgets. You only have to look at the likes of L.A. Noire (which we were not involved in!) to see how important facial animation has become for many games today.
 
In addition, the realism of facial animation has now reached the stage where the quality of the performance is becoming increasingly important.

In previous times, low fidelity facial animation would hide a relatively poor performance, but it is now becoming more and more obvious when a performance is poor.

As a result game projects are requiring better and better talent and direction. It also means that more planning is required up front, e.g. to ensure availability of key talent for a whole project.

Do you have any exciting developments in the works?
We are always continuing to develop our software and solutions and we are planning to release Version 7.0 of our DI4D software in 2013, which will have significantly improved performance and the ability to more easily handle higher density data.

However, the development we are most excited about for the year ahead will be the ability to use our software in conjunction with data acquired from head mounted camera systems. We believe that this will open up a much wider range of projects that will be able to utilise our high fidelity facial performance capture solutions.

Tell us something that no-one knows about your company.
Dug and Colin both put their heads up for sale on TurboSquid in 2003 when the company was still called Virtual Clones. Since then, Colin’s head has been purchased three times and Dug’s only once.