Rembrandt 3d printer


All Tech Considered : NPR

A Convincing Rembrandt Painting Crafted With 3-D Printing : All Tech Considered A newly unveiled portrait bearing all the hallmarks of the Dutch master is actually the result of 18 months of analysis of 346 of his paintings, plus 150 gigabytes of digitally rendered graphics.

Innovation

Heard on All Things Considered

NPR Staff

A 'New' Rembrandt: From The Frontiers Of AI And Not The Artist's Atelier

A "new" Rembrandt portrait is actually the creation of a 3-D printer — and a statistical analysis of 346 paintings by the Dutch master. Robert Harrison/J. Walter Thompson Amsterdam hide caption

toggle caption

Robert Harrison/J. Walter Thompson Amsterdam

A "new" Rembrandt portrait is actually the creation of a 3-D printer — and a statistical analysis of 346 paintings by the Dutch master.

Robert Harrison/J. Walter Thompson Amsterdam

A new Rembrandt painting unveiled in Amsterdam Tuesday has the tech world buzzing more than the art world.

That's because the painting is the creation of a 3-D printer — and not the Dutch master Rembrandt van Rijn himself, who has been dead for almost 450 years.

"The Next Rembrandt," as it's been dubbed, was the brainchild of Bas Korsten, creative director at the advertising firm J. Walter Thompson in Amsterdam.

A digital reconstruction of the face of Jesus Christ — based on skulls found in Jerusalem — inspired Korsten, he says.

"I thought, well, if you can basically take historical data and then create something new out of it, why can't we distill the artistic DNA of a painter out of his body of work and create a new artwork out of that?" Korsten tells NPR's Audie Cornish. "And that's how the idea was born."

Bas Korsten, executive creative director of the J. Walter Thompson Amsterdam agency, stands with the painting at its unveiling Tuesday in Amsterdam. Robert Harrison/J. Walter Thompson Amsterdam hide caption

toggle caption

Robert Harrison/J. Walter Thompson Amsterdam

Bas Korsten, executive creative director of the J. Walter Thompson Amsterdam agency, stands with the painting at its unveiling Tuesday in Amsterdam.

Robert Harrison/J. Walter Thompson Amsterdam

The new portrait is the product of 18 months of analysis of 346 paintings and 150 gigabytes of digitally rendered graphics.

Everything about the painting — from the subject matter (a Caucasian man between the age of 30 and 40) to his clothes (black, wide-brimmed hat, black shirt and white collar), facial hair (small mustache and goatee) and even the way his face is positioned (facing right) — was distilled from Rembrandt's body of work.

YouTube

"A computer learned, with artificial intelligence, how to re-create a new Rembrandt right eye," Korsten explains. "And we did that for all facial features, and after that, we assembled those facial features using the geometrical dimensions that Rembrandt used to use in his own work."

The statistical data even determined the type of painting — a portrait, which was the most significant portion of the artist's work, and were most common between 1632 and 1642.

The portrait looks like an actual Rembrandt, right down to the texture of the brushstrokes, which the 3-D printer mimicked. But Korsten will be the first to say that it won't fool experts.

"I wish it was that good, and it isn't," he says. "I think the expert eye sees that this isn't a real Rembrandt. And it's also got to do with the state of technology that we're in, the amount of time that we had for the process," he says. "Every extra month would have (made) a better painting."

Even a better painting will never be the same as one by the actual artist. But Korsten says the program — which started as a project for ING, the Dutch bank — could be used to help restore lost or damaged art.

"If a piece of a painting is lost, if it's burned and you're left with only 20 percent of a painting, you could, with this technique, maybe re-create the other 80 percent," he says. "So other than the advertising side of this project, I think the outcome is going to have a bigger effect, bigger impact."

Sponsor Message

Become an NPR sponsor

AI just 3D printed a brand-new Rembrandt, and it's shockingly good

News

The Next Rembrandt project used data and deep learning to produce uncanny results

Senior U.S. Correspondent, IDG News Service |

The Next Rembrandt

There's already plenty of angst out there about the prospect of jobs lost to artificial intelligence, but this week, artists got a fresh reason to be concerned.

A new "Rembrandt" painting unveiled in Amsterdam is not the work of the Dutch master Rembrandt van Rijn at all, but rather the creation of a combination of technologies including facial recognition, AI, and 3D printing.

Essentially, a deep-learning algorithm was trained on Rembrandt's 346 known paintings and then asked to produce a brand-new one replicating the artist's subject matter and style. Dubbed "The Next Rembrandt," the result is a portrait of a caucasian male, and it looks uncannily like the real thing.

One particularly interesting detail about The Next Rembrandt project, which was a collaboration among several organizations including Dutch bank ING and Microsoft, is how the algorithm chose the subject for its painting, since it had to be entirely new.

An analysis of the master's own works led the algorithm to conclude that the new piece should be a portrait featuring a Caucasian male with facial hair who is between 30 and 40 years old and is wearing dark clothing with a hat and a collar. He should also be facing toward the right, the software determined.

From there, the algorithm began to analyze specific facial features matching that profile with the aim of creating a "typical" Rembrandt eye, nose, mouth, and ear, for example. Also considered were the facial proportions, or the distances separating those features.

With the subject decided, all that remained was to bring the painting to life, and that's where 3D printing came in. The project used a 3D printer that works with a special paint-based UV ink to replicate the layers of paint and brushstrokes Rembrandt himself might have used. With 13 layers of ink, the final portrait consists of more than 148 million pixels based on 168,263 painting fragments from Rembrandt’s oeuvre.

The 18-month project was the brainchild of Dutch advertising firm J. Walter Thompson. The resulting portrait will be placed on public exhibition at some point, but details are not yet available.

Related:

  • Big Data
  • Analytics

Katherine Noyes has been an ardent geek ever since she first conquered Pyramid of Doom on an ancient TRS-80. Today she covers enterprise software in all its forms, with an emphasis on cloud computing, big data, analytics and artificial intelligence.

Copyright © 2016 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend

Rembrandt replaced with a 3D printer ?

Category:
  • Art
  • CANCEL

“The NEXT REMBRAND”

Group of specialists from Microsoft, Rembrandt House, Delfta Technical University and the Roolian Gallery of Mauritz in the APC Rembrandt" ("the Next Rebmbrandt"). When creating a portrait, machine learning technologies and 3D printing were used, which made it possible to repeat the style of the artist as much as possible, reports The Guardian.

In order to create a new painting by Rembrandt, the project team first collected data on famous paintings by the Dutch artist. The paintings were scanned in 3D, in order to record the oil relief in addition to the flowers. Using machine learning algorithms (details not disclosed), the developers improved the scanned image.

Based on the fact that Rembrandt painted many portraits, the project participants paid attention to this particular genre. When analyzing data on Rembrandt portraits, various characteristics of the depicted person were taken into account, such as age, gender, and head direction. A total of 346 portraits were processed. As a result, the following features were determined for the new portrait: it should be a white man of 30-40 years old, looking to the right, with facial hair, wearing a hat and black clothes with a white collar.

The developers, using an image recognition algorithm, have identified key details in Rembrandt's paintings, in fact, collecting a collection of eyes, mustaches, collars, facial proportions and other details of the artist's portraits. After the final image was generated, it was printed with special ink in 13 layers, for which a 3D printer was used - thus, the authors managed to convey the relief of a real portrait in oil.

Programmers, historians, art critics and data scientists were involved in the project. The final picture is based on over 168 thousand fragments of portraits painted by Rembrandt, and the project took 18 months to work on. According to the project participants, the new portrait should not be taken as an attempt to "describe Rembrandt's talent with an algorithm", it is a demonstration that data processing technologies can be used not only for business, but also for creating new works of art.

Tags: 3D printer, art

Subscribe

  • Dubai opens world's first 3D printed office The opening ceremony was attended by the ruler of Dubai Sheikh…

  • 3D-printed “Stradivari violin” was valued at six thousand euros

    3Dvarius has designed a 3D-printed electric violin. According to the creator of the tool, it is intended for professional…

  • Synthetic “nerves” work in the light

    Scheme of signal transmission along the chain of “synthetic cells”: under the action of light, genes synthesize transmembrane proteins that are integrated into . ..

Photo

Hint http://pics.livejournal .com/igrick/pic/000r1edq

AI as digital Rembrandt | Computerra

With the help of deep learning algorithms, AI was trained to copy the style of Rembrandt as accurately as possible. In the created digital portraits, even the pasty technique of the Dutch master is copied in detail. The thickness of each stroke is recreated using a 3D printer. The result looks so characteristic that at first it misleads even experts.

Digital portrait in the style of Rembrandt
(photo: fd.nl).

A digital portrait in the style of Rembrandt was presented on April 5 at the art gallery Looiersgracht 60 (Amsterdam). It consists of 148 million pixels printed in layers on a 3D printer. In addition, 13 layers of UV ink were applied at the end of the work to simulate brush marks.

The digital portrait took a total of one and a half years to complete. During this time, 346 original paintings by the master were digitized and analyzed. They studied the main plot, the type of people, their posture, head tilt, facial expression, proportions, characteristic items of clothing, the play of chiaroscuro and other distinctive features.

Examples of analyzed Rembrandt paintings (image: nextrembrandt.com).

The New Rembrandt project was financed by the Dutch financial conglomerate Ing Group. It was attended by researchers from the Delft University of Technology, the Mauritshuis Royal Gallery, the Rembrandt House Museum and Microsoft Corporation.

The total amount of primary data was about 150 gigabytes. “We studied the entire collection of Rembrandt's works, studying his paintings pixel by pixel. In order to obtain this data, we analyzed a wide range of materials, including high-resolution scans and 3D scans, which are used by deep machine learning algorithms. This extensive database was then used as the basis for The Next Rembrandt project,” commented Ron Augustus, director of Microsoft SMB marketing.

3D printer for the New Rembrandt project (photo: ING).

After statistical processing, the most typical work of Rembrandt was described as "Portrait of a European 30-40 years old with a mustache and / or beard, wearing a hat and dark clothes, head slightly tilted to the right." According to these averaged parameters, a new picture was created. The rendering itself took about 500 hours.

At the first stage, the face recognition algorithm identified and classified the most typical geometric patterns used by Rembrandt in his paintings. Machine learning principles were then used to develop his style. With their help, new facial features were generated for the future digital portrait.

Height map helps to recreate impasto technique (photo: nextrembrandt.com).

At the second stage, the vector blank was supplemented with individual features in accordance with the average proportions used by Rembrandt. The thickness of the strokes was converted into a pixel height map, which helped to recreate the characteristic texture of the canvas.


Learn more