Algorithm to Decode Ancient Texts

Scientists in Israel say they are preparing to bring online a computer algorithm that can help decode ancient texts – by looking for patterns of word choice, writing style, and even handwriting conventions. The algorithm, being put forward by researchers at Ben-Gurion University of the Negev in Beersheba, is similar to ID programs used by law enforcement officials to track and match crime scene information such as DNA.

The algorithm works its lexicographical magic by assigning a number to each pixel, depending on that pixel's darkness relative to its surroundings. Working as only a computer program can, the algorithm then, in the blink of an eye, fills in the blanks of smudges or erased text and "suggests" what might be missing. The researchers say that this kind of technology can be useful for people studying manuscripts that are faded; the researchers have gone so far as to say that they are confident that the algorithm can predict text overwritten by other subsequent hands.

So far, the researchers have focused on ancient Hebrew texts. Expansion of the algorithm to include other languages is planned.

The details of the research, already available online, will be published in the academic journal Pattern Recognition in December.