|
Automated Reconstruction and Lettering Analysis of Inscriptions
|
Topics |
3D reconstruction of ancient inscribed surfaces, segmentation and grouping of the individual letters and symbols, unbiased letter atlas construction and statistical analysis of the lettering techniques. |
|
Collaborators
Eleni Bozia, PhD
Robert S. Wagman, PhD
Visit the web-site of the DEA project:
Digital Epigraphy and Archaeology project at the UF
|
Selected Publications
A. Barmpoutis, E. Bozia, R. S. Wagman, "A novel framework for efficient 3D reconstruction and analysis of ancient inscriptions", Journal of Machine Vision and Applications, 2009, in press.
E. Bozia, A. Barmpoutis, R. S. Wagman, "An Efficient Method for Digitizing Squeezes and Performing Automated Epigraphic Analysis. ", APA 140th Annual Meeting, January 2009, Philadelphia, USA
A. Barmpoutis, E. Bozia, R. S. Wagman, "Application of 3D Technologies for the Analysis of Ancient Inscriptions.", 13th International Congress of Greek and Latin Epigraphy, September 2007, University of Oxford, UK
Click here for abstracts, BibTex entries, and PDFs.
|
Demo
| |
| This applet demonstrates our framework for 3D reconstruction of inscriptions using a shape-from-shading technique and statistical analysis of their reconstructed surfaces using images of scanned squeezes. The squeeze is scanned twice from two different directions. Using the shading information a 3D model of the original surface is reconstructed and parametrized as a height map. This map is automatically segmented into smaller regions containing single letters, which are divided into groups of same characters and are further employed for a statistical analysis of the inscribed letters. The results can be used for studying the variations of the lettering techniques within a set of inscriptions.We have applied our framework to five inscribed fragments from the archaeological site of Epidauros, containing religious hymns for Asclepius and other deities. Although found in different locations, these inscriptions show close affinities in content (sacred poetry), material (red local limestone) and writing (uncial lettering of the third century CE). The 3D reconstruction of the letterforms provides us with definite proof that all of them came from the same monument, possibly an inscribed wall from one of the buildings in the sanctuary's main area. A closer observation of the results reveals even some minor groups of letters with similar construction, which validates the accuracy of the proposed method.
|
|
|
|
|