Artomatix Raises €2.1 Million to Further Develop an Artificial Intelligence that Automatically Creates 3D, Makes Breakthrough in Style Transfer.
Dublin-based Artificial Intelligence company Artomatix Ltd. announced on Wednesday the closing of a €2.1m seed round to finance its development.
The funding round is comprised of Enterprise Ireland—an Irish government agency in charge of promoting economic development—as well as several key angel investors (€600k) and the European Commission through their Horizon 2020 SME Instrument program for high potential European companies (€1.5m).
This seed round follows €300k of pre-seed funding obtained through NDRC, a Dublin-based early-stage investor, and various grants/awards including NVIDIA Corporation’s $100k Early Stage Challenge.
This funding will be devoted to hire engineers and researchers to help execute our product roadmap and to develop our technological competitive advantage further. We believe this funding will go a long way towards helping us realize our vision to enable 3D artists create immersive worlds faster than ever.
Dr. Risser, founder and CTO
Automatically generated infinite textures: By merging traditional procedural texturing with example-based approach, Artomatix maximises variety, while minimising performance impact. The texture can then cover an area much larger than itself without displaying any repeating patterns.
Artomatix Ltd. was founded in March 2014 by Dr. Eric Risser, Neal O’Gorman, and by Barthélémy Kiss who joined shortly after. As of February 2017, the company employs 17 people.
According to Markets & Markets, 3D represents a vibrant €128 billion global industry (2016) poised for spectacular growth. By 2022, the 3D market will represent €282 billion, a 2.5x increase from current levels. Importantly, this growth is a testimonial to the central role 3D plays in numerous high-profile industries: Entertainment (Animation, Visual effects, Videogames), Virtual and Augmented reality (VR/AR), Pre-Visualization (Architecture, Online retail), Industrial design and 3D printing.
Studios creating 3D content are already at capacity and are challenged to produce the current market demands. As 3D submarkets expand, the need for 3D content increases exponentially, both in terms of quantity and quality. To address this challenge, 3D studios need to have access to services that sustainably reduce the time their 3D artists spend on creating high quality 3D assets. Otherwise this growth opportunity will come at the expense of profits.
Artomatix’s technology, known as ‘Example Based Content Creation’, relies on Dr. Risser’s expertise in Machine Learning, Computer Vision and Computer Graphics developed over ten years of research. Fully implemented, this technology will give 3D artists the ability to speed up their workflow thanks to a set of proprietary features tied to modeling—the ‘shapes’ of 3D models—and texturing—their ‘skins’.
Example-based-creation applied to 3D models: The two green heads were given as inputs to Artomatix’s technology and the blue ones were created automatically, combining the inputs features as well as inventing new ones.
‘We are convinced that the future of 3D content creation lies in Artificial Intelligence’ says Dr. Eric Risser, Artomatix’s Co-Founder & Chief Technology Officer (CTO), a Columbia University, New York and Trinity College, Dublin graduate.
A so-called ‘deep tech’ company, Artomatix develops proprietary features to speed up the 3D content creation pipeline.
In January 2017, Artomatix elevated Deep Learning-based content creation through their publication of ‘Stable and Controllable Neural Texture Synthesis and Style Transfer Using Histogram Losses.’ The company is using this technology to develop an offering that gives artists the chance to ‘recycle’ 3D assets, i.e. to breath new life into old and unusable assets by automatically updating the style and fidelity to match any need.
Style Transfer is the process of transferring the style of one picture onto the content of another. It will help 3D artists to apply the style they want to 3D environments.
To read the full text of the paper by CTO and Founder, Dr. Eric Risser as well as researchers Pierre Wilmot, and Prof. Connelly Barnes, titled ‘Stable and Controllable Neural Texture Synthesis and Style Transfer Using Histogram Losses’, follow this link.
Artomatix began a services offering in Fall 2016, geared towards the digitization of product catalogs in the industrial design industry. Artomatix counts several Fortune 500 companies among their initial clients.
Artomatix’s progress over the last three years has been boosted by recent algorithmic developments in Deep Learning as well as of the stark increase of computation power provided by Graphics Processing Units (GPUs).
Artomatix Ltd. is chaired by Dr. Steven Collins, a Partner at venture capital firm Frontline Ventures and previously the Co-Founder of Havok (a game engine middleware company), and Swrve (a mobile marketing automation platform).
Other advisors include: Dr. Stéfano Corazza, a Senior Director of Engineering at Adobe (a computer software company) and the former Founder of Mixamo (a videogame character creation platform, sold to Adobe in 2015); John Huikku, a CG Supervisor at Tau Films and a former Disney and Weta Digital artist who worked on The Lord of the Ring trilogy; Mike Walsh, a Vice-President of Developers Relations at FaceIT (a platform for online multiplayer games); and Prof. Connelly Barnes, an Assistant Professor at the University of Virginia and the inventor of Content-Aware Fill, a popular feature of Adobe’s Photoshop software.
Artomatix will be present at the Game Developers Conference (GDC) in San Francisco between February 27th and March 3rd 2017.