An art engine to take the heavy lifting out of digital art creation
Eric Risser - CTO, Artomatix
What is your idea?
Our animation creation technology uses artificial intelligence to mimic aspects of human creativity, thereby taking the heavy lifting and repetitive work out of digital art creation.
What problem are you solving and what is innovative about your approach?
Producing digital art, whether it’s for a video game or an animated TV programme or movie, is time-consuming and expensive.
Although digital art is created on computers, the process doesn't take advantage of the single most important benefit computers have to offer: automation.
Rather, art production is still largely manual, and this is partly the reason why it can cost tens of millions of dollars and take several years to produce big-name animations and games.
Our technology uses artificial intelligence based on complex mathematical equations to generate digital art iterations based on original images.
In practice, Artomatix calculates relationships between data points in an image and maps them across scale and space. It renders those data into a probability density function (PDF), which has a particular shape.
The engine can then compare, combine and manipulate similar PDFs to generate new digital images.
The process thereby harnesses human creativity but removes the need to manually generate potentially thousands of similar images that might be required, say, for initial character development or producing crowd scenes or backdrops.
So far our technology is passing the Turing test, meaning that art directors cannot tell whether art was created by human or by our program.
We are validating Artomatix technology with artists and we are in discussions with several potential customers and partners.
What’s the backstory here and how did you get involved?
I have been interested in video games and game design since I was a kid. When I was an undergraduate in the United States I started to look at ways to give computers human-like imagination, and this progressed into a Masters and PhD, during which I advanced the field of Texture Synthesis (the type of functions used now by Artomatix) by adding the concept of scaling, detecting/preserving multi-scale structures and extending a family of image-based techniques to also work on 3D models.
Artomatix was founded in early 2014, and we are based at the National Digital Research Centre as a team in the VentureLab.
Enterprise Ireland has supported our validation phase through their New Frontiers program; our plan is that we will attract private investors post NDRC VentureLab so that we would then be ready to apply for High Potential Start Up status with Enterprise Ireland.
How is this idea commercially attractive?
Through our discussions with potential customers - animation and games studios, artists and ultimately anyone who wants to create digital art - we know we are solving a huge problem for the industry.
Our animation creation engine stands to save customers money, time and effort.
The company owns and controls the IP and we are in discussions to establish a partnership within the industry that would help to promote Artomatix as a new industry standard.
Initially we are marketing our technology to bigger studios, who would purchase licenses to use the software.
Ultimately we plan to deliver Artomatix as a service that we can provide to anyone, from large-scale organisations to individual consumers who need an art item - out software can provide it.
What are you looking for at the Big Ideas event?
We are looking for investment to complete our validation of our technology and to scale and grow the company so that Artomatix becomes the new industry standard in animation creation.