Leo DORST | Hanan ELNAGHY | Michela MORTARA | Corrado PIZZI | Andreas SCALAS| Michela SPAGNUOLO
(University of Amsterdam, The Netherlands and CNR/IMATI Genova, Italy, Amsterdam, The Netherlands)

Keywords: mesh processing, multiple resolution, mesh segmentation, annotation transfer

In the GRAVITATE project, we process semantic and geometric information about scanned artefacts, to provide digital tools for the workflow of archaeological analysis. In this talk, we will focus on the 3-part ‘geometric ingestion’ pipeline that prepares 3D digital models of artefacts for treatment by our geometric algorithms. Resampling [1] – Archaeological meshes from a high resolution scanner may look fine, but at the vertex level typically contain many defects, such as holes, connectivity errors or intersecting triangles. So we developed an automatic cleaning and simplification procedure, which fixes geometry and topology flaws and simplifies models while maintaining surface details. Faceting [2] – The geometric tools in GRAVITATE attempt to perform reassembly and/or compute geometric properties of the original outside of the broken fragment to characterize them for search. They therefore need a mesh that is segmented into ‘fractures’ and ‘skin’ sub-meshes. We call the pre-processing which discriminates those zones faceting. We developed a robust algorithm for faceting, with few tuneable yet archaeologically meaningful parameters. Annotation Transfer [3] – In the GRAVITATE system, each fragment is represented by meshes of different resolutions; the main reasons are efficiency in the processing (if some aspects can be adequately resolved at low resolution; e.g. faceting runs on 50k meshes) and interactive display (such as in a fast web interface). These different versions need to be kept consistent, not only as meshes, but also with their GRAVITATE annotations (like faceting, or sub-feature labelling). We present an annotation transfer procedure which projects facets from one resolution to another so that further downstream tools (with different resolution requirements) can take advantage of them transparently. The transfer can handle any annotation, i.e., a part of the geometry with specific associated semantics; we believe this practice is crucial to fostering new strategies for CH documentation.

Relevance for the conference: Describes crucial issues involved in handling larage 3D meshes for archaeology
Relevance for the session: This is how we approached those issues in GRAVITATE
Innovation: Faceting with intuitive parameters and annotation transfer across resolutions
• Mortara, M., C. Pizzi, and M. Spagnuolo. 2017. “Streamlining the Preparation of Scanned 3D Artifacts to Support Digital Analysis and Processing: the GRAVITATE Case Study.” Proceedings of GCH 2017 – Eurographics Workshop on Graphics and Cultural Heritage. Eurographics. ISSN 2312-6124, ISBN 978-3-03868-037-6, DOI 10.2312/gch.20171309
• ElNaghy, H., and L Dorst. 2017. “Geometry Based Faceting of 3D Digitized Archaeological Fragments.” ICCV e-Heritage Workshop. Venice: ICCV. 2934-2942. pdf.
• Scalas, A., M. Mortara, and M. Spagnuolo. 2017. “3D Annotation Transfer.” Proceedings of GCH 2017 – Eurographics Workshop on Graphics and Cultural Heritage. The Eurographics Association. ISSN 2312-6124, ISBN 978-3-03868-037-6, DOI 10.2312/gch.20171295