Proceedings of the 2011 Workshop on Historical Document Imaging and Processing
Thousands of documents written in Syriac script by early Christian theologians are of unknown provenance and uncertain date, partly due to a shortage of human expertise. This paper addresses the problem of attribution by developing a novel algorithm for offline handwriting style identification and document retrieval, demonstrated on a set of documents in the Estrangelo variant of Syriac writing. The method employs a feature vector based upon the estimated affine transformation of actual observed characters, character parts, and voids within characters as compared to a hypothetical average or ideal form. Experiments on seventy-six pages from nineteen Syriac manuscripts written by different scribes show that the method can identify pages written in the same hand with high precision, even with documents that exhibit various challenging forms of degradation.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
© the authors
Dalton, Emma and Howe, Nicholas, "Style-Based Retrieval for Ancient Syriac Manuscripts" (2011). Computer Science: Faculty Publications, Smith College, Northampton, MA.