A new approach for ancient inscriptions' writer identification

Rousopoulos, P; Panagopoulos, M; Papaodysseus, C; Panopoulou, F; Arabadjis, D; Tracy, S; Giannopoulos, F; Zannos, S

dc.contributor.author	Rousopoulos, P	en
dc.contributor.author	Panagopoulos, M	en
dc.contributor.author	Papaodysseus, C	en
dc.contributor.author	Panopoulou, F	en
dc.contributor.author	Arabadjis, D	en
dc.contributor.author	Tracy, S	en
dc.contributor.author	Giannopoulos, F	en
dc.contributor.author	Zannos, S	en
dc.date.accessioned	2014-03-01T02:47:14Z
dc.date.available	2014-03-01T02:47:14Z
dc.date.issued	2011	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/33027
dc.subject	ancient inscriptions dating	en
dc.subject	automatic writer identification	en
dc.subject	curve fitting	en
dc.subject	graphological pattern recognition	en
dc.subject.other	Alphabet symbols	en
dc.subject.other	ancient inscriptions dating	en
dc.subject.other	Archaeometry	en
dc.subject.other	Goodness of fit	en
dc.subject.other	Objective methods	en
dc.subject.other	Pairwise matching	en
dc.subject.other	Statistical criterion	en
dc.subject.other	Statistical distribution	en
dc.subject.other	Writer identification	en
dc.subject.other	Digital signal processing	en
dc.subject.other	Image processing	en
dc.subject.other	Pattern recognition	en
dc.subject.other	Curve fitting	en
dc.title	A new approach for ancient inscriptions' writer identification	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/ICDSP.2011.6004966	en
heal.identifier.secondary	http://dx.doi.org/10.1109/ICDSP.2011.6004966	en
heal.identifier.secondary	6004966	en
heal.publicationDate	2011	en
heal.abstract	A novel approach is introduced here for identifying the writer of an ancient inscription, via of methods of pattern recognition, image processing and mathematics. Identifying the writer of an ancient inscription is of fundamental importance for Archaeometry and History, since, it is, so far, the only objective method for dating its content. The introduced method consists in a) pairwise matching of all realizations of an alphabet symbol on two inscriptions. This matching is performed so as to compensate for systematic differences in size, orientation and position of the various letters of the documents, b) a new criterion that quantitatively describes the goodness of fit is given, c) the statistical distribution of quantities associated with this criterion, are estimated, d) novel statistical criteria for deciding if two inscriptions are made by the same writer or not, have been developed. The approach has been successfully applied to 32 Classical and Hellenistic inscriptions, flawlessly classifying them into 9 writers. © 2011 IEEE.	en
heal.journalName	17th DSP 2011 International Conference on Digital Signal Processing, Proceedings	en
dc.identifier.doi	10.1109/ICDSP.2011.6004966	en