Tech start-up of the week: Artomatix

19 Apr 2014

Dr Eric Risser and Neal O'Gorman, founders of Artomatix

Our start-up of the week is Artomatix, whose team has built an artificial intelligence system with human-like artistic creativity.

Artomatix has kicked off a Kickstarter-style crowdfunding campaign on Trinity College’s new crowdfunding page.

The technology works in such as way that once artists supply Artomatix Studio with sample 2D or 3D artwork, the software then generates new art based on these inputs.

“Our vision is that all digital graphic artists will use our technology but initially we are focused on the movie and video-game industries, which we calculate to be a total addressable market of at least €225m,” explained co-founder Dr Eric Risser.

“When you include all other digital graphic artists, the total opportunity sums up to greater than €500m.”

At the moment, Risser and his colleague Neal O’Gorman are working to release a standalone application for high-end desktops which will sit alongside professional artist software, such as Adobe’s Photoshop, Autodesk’s 3D Studio Max/Maya, and Pixologic’s ZBrush, becoming a crucial pillar for most video-game and movie production.

“Our technology is based on 20 years of academic research … From a business standpoint we’re solving the single largest pain of both the video game and movie industries – art creation takes too long and costs too much!”

The founders

Risser is an expert in the combined fields of artificial intelligence and computer graphics and the pioneer of the Artomatix technology. He completed his master’s degree at Columbia University and PhD at Trinity College Dublin (TCD).

“I have authored six technical publications during my academic career at Columbia University and Trinity College Dublin. I have given talks at top industry/academic conferences, such as Game Developers Conference (GDC) and Siggraph, multiple times each. I’ve also been invited to speak at a number of companies/institutions, including Pixar and Princeton University.”

His co-founder O’Gorman is a serial entrepreneur, with Artomatix being the third company he has founded – the first was acquired by Agilent Technologies in 2006.

His second company built a social game that had an acquisition offer from one of the top social games companies in the world. O’Gorman has a technical background with three patents to his name but in the last decade has been more commercially focused, fulfilling many roles such as CEO, product manager, portfolio manager, sales and business development.

“Neal and I first met each other in November 2012 through NDRC’s pre-accelerator called Seedlab. This gave me an opportunity to progress customer discovery and customer validation while at the same time work with experienced entrepreneurs like Neal. In January 2013, I was delighted when Neal accepted my offer to come onboard as a co-founder in Artomatix.

“Our advisers and mentors include founders of high-profile game middleware companies, including Havok and Demonware, as well as other leading scientists in this field,” Risser said.

The technology

“At a superficial level, our technology works kind of like a spreadsheet does,” Risser explains. “The same way an accountant would enter meaningful high-level data and the spreadsheet does all the tedious calculations, our system lets an artist enter their concepts as input and our system helps them quickly polish and improve those concepts, and then build a large virtual world with a level of detail and variety that mirrors the real world.

“Under the hood our technology is very sophisticated. It’s been a topic of academic research for over 20 years, with myself and our adviser Prof Connelly Barnes being two of the few experts in the world on the subject. It’s a fully ‘example based’ approach where you give the computer a sample of art you like, it learns it and then can produce new things by using your example as a template.

“At its core our technology spans several fields of study including computer vision, machine learning and graphics. The learning/creating strategy is similar to the general ‘big data’ algorithm where deep understanding is made unnecessary through the use of pattern matching.

“In general, the first thing our system learns to do is match and correlate simple patterns. It then begins grouping and combining those simple patterns into higher-order more sophisticated patterns. By treating a piece of art as a collection of patterns, we can find connections and build links between artistic features in very different inputs.

“This is similar to the way Google Translate works where they exposed it to EU legal documents where one single message must be translated into every European language, this allows them not just to connect words, but entire sentences and phrases. Our problem is a bit more complicated than translation though, after we’ve figured out the way patterns connect, we still have to imagine new things, new things that still conform to the underlying patterns. That’s where the magic comes in to play.”

Transforming the video-game and movie industries

Risser said the ultimate goal of the company is to transform the video-game and movie industries by giving them better and smarter tools.

“Right now there’s no way to make lots of art cheaply, in recent years this has caused a huge class divide in the industry. Constantly rising production costs have caused the old-school video game/movie companies from the Nineties and early 2000s to either merge together or die off.

“The only projects that are economically viable these days are smash hit AAA blockbusters, either sequels or based off of some already popular brand. Producers in this space have vast resources, but no artistic/creative freedom. At the other end of the spectrum there’s a bloom in activity.

“In the casual games/YouTube space, small projects with little to no budget can still be very commercially viable if they’re cool and unique. Producers in this space have full artistic/creative freedom, but are limited by their lack of resources. It’s clear that projects are currently being dictated by economic niche, rather than creative vision. This makes the industry poorer as a whole.

“The ultimate goal is to fill in the chasm between high and low end projects, remove the class divide in the industry and enable video-game and movie companies to make bigger better projects faster and cheaper.”

Turning coal into diamonds

Risser said Artomatix is getting an overwhelmingly positive response from its customers.

“We’ve recently started our Alpha testing phase and are about to initiate the Beta. The plan is to have a v1.0 release in Q4 of this year.”

The company was initially bootstrapped through the Launchbox student incubator programme at Trinity College before being selected by Enterprise Ireland for its New Frontiers programme

“This has offered us more capital and professional mentorship and more importantly a channel into the entrepreneurial scene in Dublin and all the connections that can provide. We’ve just been accepted into NDRC’s VentureLab accelerator and with their investment of €100,000 we’re looking to hire more programmers and grow.

“We’re close to starting fundraising, and have identified the investors in our space internationally, as well as having begun early conversations with local investors, both VC and angel.”

Risser said the hardest part was getting started. “There’s a huge chasm between the academic and entrepreneurial worlds, in terms of mindset, culture and networks. In fact being an academic or an entrepreneur makes you an outlier that few tasks in everyday life would prepare you for.

“In academia you solve problems that are difficult. As an entrepreneur you solve problems that are important. While these two characteristics aren’t necessarily exclusive from each other, the mindset and optimal approach to identifying and solving them are completely different and tends to favour one or the other. As an example, turning coal into diamonds is an excellent problem for an entrepreneur while turning diamonds into coal would be an excellent problem for an academic.”

It’s a unique and exciting time to be a founder in Ireland, Risser said. “Due to various economic and political forces, there are many support systems in place for first timers. Our advice is to avail of them.

“No matter what your venture is, be it a new high-tech invention, a business model innovation on a web service, or even introducing a new type of health food to the Irish market, there’s a private or public support system to help you succeed at it.”

John Kennedy is a journalist who served as editor of Silicon Republic for 17 years

editorial@siliconrepublic.com