CNN  — 

A nude portrait of a crouching woman, hidden beneath the surface of a Pablo Picasso painting, has been revealed using artificial intelligence, advanced imaging technology and 3D printing.

Dubbed “The Lonesome Crouching Nude,” the recreation is the work of Oxia Palus, a company that uses technology to resurrect lost art, according to a statement sent to CNN on Monday.

Picasso painted over the figure when making “The Blind Man’s Meal” in 1903. The nude had been partially revealed by a superimposed X-ray fluorescence (XRF) image, but Oxia Palus has now “brought the hidden work to life back to life,” according to the statement.

In order to do so, the company used XRF imaging and image processing to reveal the outline of the hidden painting, and then trained artificial intelligence to add brushstrokes to the portrait in the style of Picasso.

The Metropolitan Museum of Art
Picasso painted over the nude to make "The Blind Man's Meal" (pictured) in 1903.

It then generated a heightmap of the portrait to give it texture, and printed the image onto canvas using 3D printing technology.

Oxia Palus was co-founded by George Cann and Anthony Bourached, a pair of PhD candidates in machine learning at University College London (UCL).

Art is a store of complex information, and machine learning has developed to the point that it can help us analyze that information, Bourached told CNN.

“We’ve got complex systems now that can help us understand our history and culture better,” he said.

While X-ray images are useful in revealing images that were painted over, the AI adds another layer to our analysis, Bourached added.

The nude dates from Picasso’s Blue Period, early in his career, and the artist wouldn’t have wanted to paint over the figure, according to Cann.

The Metropolitan Museum of Art
An X-ray of the "The Blind Man's Meal" partially revealed the portrait.

“At the time that Picasso painted The Lonesome Crouching Nude and The Blind Man’s Meal he was poor and artist materials were expensive, so he likely painted over the former work with reluctance,” Cann said in the statement.

The figure of the woman also appears in Picasso’s painting “La Vie” and a number of his sketches, which suggests the artist may have had an affinity for her, he added.

“I hope that Picasso would be happy in knowing the treasure he’s hidden for future generations is finally being revealed, 48 years after his death and 118 years after the painting was concealed,” Cann said in the statement.

Ty Murphy, a Picasso specialist at London-based due-diligence firm Domos Art Advisors, said the AI-produced painting looks like a Picasso Blue Period, but under close scrutiny an expert would probably be able to recognize that it wasn’t an original.

However, continued developments in machine learning and 3D printing should make for even more accurate work in the future, he added.

“Give it time,” said Murphy: “As soon as those technologies emerge they’re going to be very convincing.”

While some in the art world have criticized the approach, Murphy has no issue with using machine learning to produce new works.

“History has shown us that people will always emulate the work of other artists,” he said, adding: “It’s an exploration of Picasso’s mind.”

David Dibosa, who leads the Masters course in Curating and Collections at Chelsea College of Arts in London, said anyone who enjoyed Picasso’s work would be excited by the emergence of the image, and hailed the combination of technology at work in the piece.

However, he questioned whether there was any need to print the portrait on canvas, asking whether it would have been more accessible if it were a digital piece.

“There are many people who would never be able to see a limited edition artwork, even if it were shown in a museum nearby,” said Dibosa.

“By highlighting The Lonesome Crouching Nude as a digital discovery made for an online world,” he said, “we could bring Picasso front and center to meet the new realities of the 21st century.”

The artwork will be on display at the Deeep AI Art Fair in London from Wednesday to Sunday.