News - Digital Restoration Initiative

In the News

Dr. Seales on 60 Minutes.

Virtual Unwrapping and DRI Research Featured in International Media

The work of Professor Brent Seales and his Digital Restoration Team were featured recently on the PBS science show NOVA and on the streaming network, Curiosity Stream. The hour-long NOVA episode aired November 6th. Curiosity Stream is a subscription-based, science-focused streaming network from the founder of Discovery Channel. Two clips from the excellent show can be viewed here and here, but users can sign up for a free two week trial to view the entire 20-minute Breakthrough segment, called “Herculaneum Scrolls: Unravelling History.”

The team’s recent scans of Herculaneum scrolls at Diamond Light Source, a synchrotron located just outside of Oxford, England, garnered worldwide attention, including an appearance on BBC Radio’s Inside Science. (See all of the international press coverage here.) A TV production team from the U.K. also followed Seales and his team to Diamond to record the scan session and create this “teaser” video for a documentary about Seales and his work. The final production will appear on networks around the world, including NatGeo in the U.K. and SBS in Australia.

Prof. Brent Seales Using Light Brighter Than the Sun to Read Herculaneum Scrolls

Using light 10 billion times brighter than the sun, Professor Brent Seales and his team recently captured images of authentic Herculaneum material at Diamond Light Source, a high-energy physics facility in Oxford, England. Thanks to funding from the National Endowment for the Humanities and the Mellon Foundation, the team was able to purchase the amount of beam time needed to scan — at unprecedented spatial resolution — two intact scrolls and four fragments belonging to the Institut de France in Paris.

“This scan session promises to be a key moment in our quest for a reliable pathway to reading the invisible library,” said Seales. “The brightness and speed of the Diamond beam, and its ability to handle massive data sets, are crucial. We are using the Diamond facility to acquire scans of complete scrolls and also open fragments, with the hope that the open fragments will form a reference library — or training set — that can inform our machine learning software approach to visualizing carbon ink,” Seales continued. “If things work as we expect, the scans and the machine learning will open up the chance to see the ink very clearly for the first time.”

Watch video

Brent Seales - $2 million grant from The Andrew W. Mellon Foundation

DRI Receives $2 Million Mellon Foundation Grant

The University of Kentucky is poised to become a world-class leader in “unwrapping” cultural artifacts. Thanks, in large part, to a $2 million grant from The Andrew W. Mellon Foundation, Professor Brent Seales finally has the materials access, funding support and technical approach needed to solve the 2,000-year-old mystery of the Herculaneum papyri.

The prestigious Mellon grant will provide the resources the team needs to virtually unwrap and digitally restore the scrolls. It will also support the electronic compilation and dissemination of the entire Herculaneum collection, which is currently spread across four different institutions: the Bodleian Library at Oxford University, the British Library, the Institut de France and the Biblioteca Nazoinale di Napoli.

Innovative Machine Learning Tool Reveals Carbon Ink

A new study by Dr. Seales and his team refutes claims that carbon ink is “invisible” in micro-CT scanning and conclusively shows that machine learning can be used to elicit the stubborn text. Their groundbreaking work, featured in the science e-journal PLOS One (From Invisibility to Readability: Recovering the Ink of Herculaneum) demonstrates how characteristics other than density differences, which is the traditional imaging factor in x-ray, can be capitalized upon in micro-CT to reveal carbon ink writing. The team developed a 3D Convolutional Neural Network (CNN) that identifies the unique data pattern generated by the scanner when ink is present on the surface of papyrus, versus when the papyrus is blank. In the study, the team successfully applied the CNN to both proxy materials and the most well-known and sought-after of carbon-inked texts, Herculaneum papyri. The team also developed a modification of the tool that can generate color renditions from the scans of the inked papyrus, thus providing a realistic facsimile of the original instead of just a traditional black and white x-ray image.

‘Behind the Blue’: UK’s Brent Seales Digitally ‘Unwraps’ Ancient Texts, Other Treasures

By Carl Nathe

LEXINGTON, Ky. (May 15, 2019) — Thanks to work spearheaded by University of Kentucky College of Engineering faculty member W. Brent Seales, UK is poised to become the world-class leader in digitally “unwrapping” and restoring one-of-a-kind cultural artifacts, such as ancient manuscripts.

Seales and his students have worked for more than two decades to noninvasively image and unfurl all types of fragile texts, such as “Beowulf,” the Dead Sea Scrolls and more. Seales, professor and chair of UK’s Department of Computer Science, has earned a reputation as “the man who can read the unreadable.”

PI Professor Brent Seales Awarded NEH Grant to Uncover Texts of Herculaneum

By Lindsey Piercy March 5, 2019

Professor Brent Seales has received a National Endowment for the Humanities (NEH) grant for his groundbreaking project, “Reading the Invisible Library: Rescuing the Hidden Texts of Herculaneum.” The Digital Humanities Advancement Grant of $325,000 will allow Seales and his dedicated team to continue development of computerized techniques to recover writings from the Herculaneum library, a collection of undecipherable papyrus scrolls were carbonized during the eruption of Mount Vesuvius in 79 CE (Common Era). More specifically, funding from the NEH is being used to develop a machine learning approach that will enable researchers to see hidden writing that is otherwise very difficult to visualize in X-ray based images. The funds support the construction of a large-scale neural network, including student and staff time for software development.