DeepScribe: Localization and Classification of Elamite Cuneiform Signs via Deep Learning

Twenty-five hundred years ago, the “paperwork” of the Achaemenid Empire was recorded on clay tablets. In 1933, archaeologists from the University of Chicago’s Institute for the Study of Ancient Cultures (ISAC, formerly Oriental Institute) found tens of thousands of these tablets and fragments during the excavation of Persepolis. Many of these tablets have been painstakingly photographed and annotated by expert cuneiformists, and now provide a rich dataset consisting of over 5,000 annotated tablet images and 100,000 cuneiform sign bounding boxes encoding the Elamite language. We leverage this dataset to develop DeepScribe, the first computer vision pipeline capable of localizing Elamite cuneiform signs and providing suggestions for the identity of each sign. We investigate the difficulty of learning subtasks relevant to Elamite cuneiform tablet transcription on ground-truth data, finding that a RetinaNet object detector achieves a localization mAP of 0.78 and a ResNet classifier achieves a top-5 sign classification accuracy of 0.89. The end-to-end pipeline achieves a top-5 classification accuracy of 0.80. As part of the classification module, DeepScribe groups cuneiform signs into morphological clusters. We consider how this automatic clustering approach differs from the organization of standard, printed sign lists and what we learn from it. These components, trained individually, are sufficient to produce a system that can analyze photos of cuneiform tablets from the Achaemenid period and provide useful transliteration suggestions to researchers. We evaluate the model’s end-to-end performance on locating and classifying signs, providing a roadmap to a linguistically aware transliteration system, then consider the model’s potential utility when applied to other periods of cuneiform writing.