Computer scientists at Princeton University in New Jersey have developed an automated system that could speed a painfully slow task facing archaeologists – reconstructing ancient artifacts from thousands of tiny, often-eroded pieces.David Dobkin, computer science professor and dean of the faculty at Princeton, got the inspiration for the project after visiting the archaeological site of Akrotiri on the island of Thera, which in modern-day Greece is known as Santorini.There the archaeologists have been attempting the Herculean task of reconstructing wall paintings that hold valuable clues to the ancient culture, an island civilization buried under volcanic ash more than 3500 years ago. The archaeologists have been working on the project for several decades, but at their current rate, it would still take more than a century to complete their fresco fragment puzzle.The team’s approach involves two scanners: an ordinary flatbed document scanner that scans the colors of the front side of the fragments, and a laser scanner that scans the geometry, or shape, of the fragments while also acquiring some rudimentary color information with its built-in camera. A motorized turntable connected to the laser scanner allows for precise rotation of the fragment as it is being studied, and both scanners are connected to a laptop computer.A 3-D laser scanner setup that includes a motorized turntable. Currently deployed at the fresco laboratory at the Akrotiri excavation.The new technology “has the potential to change the way people do archaeology,” Dobkin said. “This approach really brings in the computer as a research partner to archaeologists.”The ancient artifacts excavated at Akrotiri, which flourished in the late Bronze Age around 1630 before the common era, are often thousands of tiny, eroded pieces of a fresco, mosaic or similar object. Many lack any distinctive color, pattern or texture and have had their edges eroded over the centuries, making it very difficult to know how they fit together. Sifting through the fragments, archaeologists find matches mainly through trial and error.While other researchers have developed computer systems to automate parts of the process, their approaches often rely on large pieces of expensive equipment that must be operated by trained computer experts.The Princeton technique uses inexpensive, off-the-shelf hardware and is designed to be operated by archaeologists and conservators in the field. “We mimic the archaeologists’ methods as much as possible, so that they can really use our system as a tool,” said Szymon Rusinkiewicz, an associate professor of computer science at Princeton. “When fully developed, this system could reduce the time needed to reconstruct a wall from years to months. It could free up archaeologists for other valuable tasks such as restoration and ethnographic study.” A large Princeton team made several visits to Akrotiri in 2007 to observe, learn and later test the system. During a three-day visit to the island in September 2007, they used it to successfully measure 150 fragments.Although the system is still being perfected, it already has yielded promising results on real-world samples. For instance, when tested on a subset of fragments from a large Akrotiri wall painting, it found 10 out of 12 known matches. It also found two additional matches that were previously unknown. Above, Pottery fragments from the Akrotiri excavation at Thera, Greece, which were matched using a system devised by Princeton University computer scientists. Archaeologists can produce complete 3-D models of at least 10 fragments per hour using the system, which includes a flatbed scanner and a 3-D laser scanner. To efficiently compute fragment matches, their edges are re-sampled by the computer into a mesh called a “ribbon.” These six sherd matches from the spiral data set were found with the ribbon matcher; the decoration’s continuity just confirms success. Images courtesy of Princeton Graphics Group. “This showed that the system could work in a real-life situation,” said Tim Weyrich, the technical lead researcher on the project who designed many of its components. Formerly a postdoctoral teaching fellow in computer science at Princeton, he is now an assistant professor in the computer science department at University College London. By following a precisely defined and intuitive sequence of actions, a conservator working under the direction of an archaeologist can use the system to measure, or “acquire,” up to 10 fragments an hour. The flatbed scanner first is used to record several high-resolution color images of the fragment. Next, the fragment is placed on the turntable, and the laser scanner measures its visible surface from various viewpoints. The fragment is then turned upside down, and the process is repeated. The software, dubbed “Griphos,” which is Greek for puzzle or riddle, processes all the information. One algorithm aligns the various partial surface measurements to create a complete and accurate 3-D image of the piece. Another analyzes the scanned images to detect cracks or other minute surface markings that the laser scanner might have missed. The system then integrates all of the information gathered – shape, image and surface detail – into a meticulous record of each fragment. “This in itself is extremely useful information for archaeologists,” Weyrich said. “Our initial evaluation was greatly received, so we permanently installed a scanning station at Akrotiri, which took place during a trip in September.” Once it has acquired an object’s fragments, the system begins to reassemble them, examining a pair at a time. Using only the information from edge surfaces, it acts as a virtual archaeologist, sorting through the fragments to see which ones fit snugly together. Analyzing a typical pair of fragments to see whether they match is very fast, taking only a second or two. However, the time needed to reassemble a large fresco may be significant, because the system must examine all possible pairs of fragments. To make the system run faster, the researchers are planning to incorporate a number of additional cues that archaeologists typically use to simplify their searching for matching fragments. This data includes information such as where fragments were found, their pigment texture and their state of preservation. The scientists emphasize that the system is meant to assist, not replace, conservators and archaeologists.“Reconstructing these frescoes is incredibly complex, given the condition of the fragments and the sheer number of fragments,” Weyrich said. “The computer takes over the laborious parts of the process while leaving the important, intuitive decisions to the humans.”