Goal: When completed this tool will allow researchers to take a picture of a cunieform tablet and have it transliterated in C-ATF.
There are thousands of cuneiform tablets stored away in various collections around the world, but so few people skilled enough to translate them. Meaning the vast majority of discovered tablets are just locked away collecting dust. The Cuneiform Digital Library Initiative(CDLI) is one of the few programs attempting to combat this, by providing a public archive of professional cuneiform research, collections, and resources to aide researchers and amatuers about the state of cuneiform language and history.
Part 1: Use Deep Learning(possibly a variation of YOLO algorithm) to develop a model that can detect all of the cuneiform symbols present in an image of cuneiform tablet.
Part 2: Use DL to develop a model to identify each symbol.
Part 3: Use an existing api(cuneify) to encode each symbol to C-ATF.
Once completed this tool will allow researchers swiftly transliterate tablets and enter them into the CDLI catalogue. Thus making thousands of previously inaccessible tablets available to the greater research community.
Cuneiform script, one of the earliest systems of writing, was invented by the Sumerians. It is distinguished by its wedge-shaped marks on clay tablets, made by means of a blunt reed for a stylus. The name cuneiform itself simply means "wedge shaped".
Cuneiform is not a language, nor is it an alphabet. Cuneiform uses between 600-1000 characters to write words or syllables. It has been used by many different cultural groups to represent many different languages, but it was primarily used to write Sumerian and Akkadian. Deciphering cuneiform is very difficult to this day, though the difficulty varies depending on the language.