Watch This AI Turn Sketches into Masterpieces
Trained on the ‘history of human art,’ this system can transform your scribbles
/https://tf-cmsv2-smithsonianmag-media.s3.amazonaws.com/filer/6f/e4/6fe4e332-9e90-44e7-bd44-ff7e5ae39d89/maxresdefault.jpg)
As the field of artificial intelligence has advanced rapidly in recent years, engineers have harnessed its power to create trippy art and music, generate wonky fake disease names and even write a Game of Thrones sequel. Now, a newly developed system can turn your sketches into a masterpiece influenced by the sum of the world's greatest artists.
“What we’ve built would have been unthinkable to the original deep learning pioneers,” said Monty Barlow, director of machine learning at Cambridge Consultants, in a statement. "We’ve created something hugely interactive, taking the germ of a sketched idea and allowing the history of human art to run with it.”
Barlow's team has created "Vincent," an AI that can take a person's simple drawings with a stylus on a screen and build on them to make them make a finished work with color, a background and texture.
Vincent's creators "trained" the system by getting it to study thousands of works of art dating back to the Renaissance era to give it an understanding of how the techniques of painting work and look, reports Steve Dent for Engadget. This machine learning enables the algorithms of Vincent to take a person's drawing and respond to it by fleshing the work out into something resembling an actual painting.
"Unlike typical machine learning approaches which simply use mathematics to generate approximations of art, Vincent is able to take the human input given and produce a relevant, finished artwork," the company said in a statement.
Vincent may just be for fun, but its creators envision the systems behind it having practical applications in other fields. AI like this could be used to create scores of unique training scenarios for autonomous vehicles or digital security based on simple input from humans, saving lots of work. Barlow's team, which unveiled Vincent at a summit in London last week, is looking forward to pushing the envelope further in this burgeoning field.
“We’re exploring completely uncharted territory–much of what makes Vincent tick was not known to the machine learning community just a year ago,” Barlow said.
