Feldmann, FelixBogacz, BartoszPrager, ChristianMara, HubertTobias Schreck and Tim Weyrich and Robert Sablatnig and Benjamin Stular2017-09-272017-09-272017978-3-03868-037-62312-6124https://doi.org/10.2312/gch.20171301https://diglib.eg.org:443/handle/10.2312/gch20171301Deciphering the Maya writing is an ongoing effort that has already started in the early 19th century. Inexpertly-created drawings of Maya writing systems resulted in a large number of misinterpretations concerning the contents of these glyphs. As a consequence, the decryption of Maya writing systems has experienced several setbacks. Modern research in the domain of cultural heritage requires a maximum amount of precision in capturing and analyzing artifacts so that scholars can work on - preferably - unmodified data as much as possible. This work presents an approach to Maya glyph retrieval based on a machine learning pipeline. A Support Vector Machine (SVM) classifier is trained based on the Histogram of Oriented Gradients (HOG) feature descriptors of the query glyph and random background image patches. Then a sliding window classifies regions into viable candidates on the scale pyramid of the document image to achieve scale invariance. The algorithm is demonstrated on two different data sets. First, photographs from a hand written codex and second 3D scans from stone engraved monuments. A large amount of future extensions lies ahead, comprising the extension to 3D, but also more sophisticated classification algorithms.Computing methodologiesShape representationsObject identificationApplied computingGraphics recognition and interpretationOptical character recognition10.2312/gch.20171301115-118