Computer attributes stone tablet writing to specific author


Overkill VR is a fast paced VR action game

A computer system has been developed that can differentiate between various authors on stone tablets from ancient Greece. Two computer engineers at the National Technical University of Athens, Michail Panagopoulos and Constantin Papaodysseus, worked on the project to create a system for writer identification which they have applied to ancient Greek inscriptions.

Their group has focused on several projects which include computer system design and analysis, pattern recognition, and image processing. Their research involves virtual reconstruction of fragmented objects, writer identification and graphological examination of ancient artifacts, automatic recognition of musical recordings and video streams, and computer algorithms.

Their recent work is assisting archaeologists place fragments, and thus the information that can be gleaned from them, in their proper historical sequence. Linking the writing style on one stone tablet to that on another helps archaeologists put a time line on artifacts they are studying.

Credit ? Athens Epigraphical Museum

Stephen Tracy, a Greek scholar and epigrapher at the Institute for Advanced Study in Princeton, New Jersey, has spent years studying stone tablets. You might say he a handwriting analysis expert, although he is not using it to interpret the personality of the writer. Instead, his concentrated studies have honed his skill so that he can tell one writer from another.

The analysis was done over six distinct Greek letters  ?, ?, ?, ?, ? and ?. They looked at 24 pieces from tablets ranging in age from 334 BC to134 BC, attributed to six different artists who carved the inscriptions. Tracy said the computer results were 100% correct.

In the future, for greater accuracy, the researchers hope to use three-dimensional laser scans in their analysis, instead of digital photographs. They also want to build a data base against which future discoveries can be compared.

The translation of the image accompanying this article, a 357 BC stele from Thrace can be found at the Institute for Advanced Studies? website. The researchers? study appears in IEEE Xplore: IEEE Transactions on Pattern Analysis and Machine Intelligence, published in the last editions of the New Scientist and the American Journal of Archaeology.