Computer games techniques for optimizing acquired 3D models in museums applications

Paper
Gabriele Guidi, Italia

Since the introduction of 3D acquisition in Cultural Heritage (CH) in the very early experiments of the late ‘90, it was clear that one of the most important bottlenecks in the application of 3D models for museum applications would have been the huge size of the models in term of polygons, and consequently the huge size of the related files (from hundreds of MegaBytes to a few GigaBytes), whose management and visualization was not feasible without custom computers and highly specialized software.

Nowadays, at more than 15 years by the Digital Michelangelo Project, although several new tools have been developed, the main obstacle to a widespread diffusion of 3D technologies in museums is still the overwhelming difference between the easiness of collecting plenty of 3D data with laser scanners or computer vision techniques, that allows to create extremely detailed 3D contents potentially very suitable for engaging interactions in museum applications, and the still difficult management of huge 3D polygonal meshes with everyday computers.

In the last few years the powerful 3D technologies developed for the highly dynamic 3D games nowadays available on the market, have been demonstrated to be usable for implementing “serious gaming” powerful digital applications for CH suggesting interesting potentiality specifically for museums. In this framework the step still under development is the transformation of the detailed 3D models coming out from a 3D acquisition process, to a 3D model suitable for a gaming platform, that has specific characteristics in terms of size and topological constraints. Although several tools has been developed for the gaming industry, such transformation is still not straightforward for complex acquired models.

This papers illustrated a possible pipeline for solving such step with open source or off-the-shelf software tools, demonstrating the results on a set of 3D models acquired at the Archaeological Museum of Milan, in the framework of the 3D-Icons European Project (http://3dicons-project.eu), aiming at generating about 3000 new 3D models and the related metadata for populating the Europeana portal. The main idea behind this transformation is not new. It’s based on reducing a complex mesh to a very simplified 3D model giving a rough approximation of the actual geometry, plus a map of deviations for compensating the lack of information of the 3D object, using a Displacement Mapping (DM) technique. DM actually operates on the geometry rather than just at visual level as bump, normal or parallax mapping, moving 3D vertices of a smooth surface in order to add finer geometrical details. This paper shows a pipeline for transforming a complex texturized mesh model in a subdivision surface (Sub-D) plus a displacement map extracted from the acquired 3D model. The pipeline is highly automatic, leaving a very limited set of manual actions to the operator. The derived 3D model in addition to a huge size reduction, is characterized by a structure easily portable to a 3D game engine of the last generation. Examples of archaeological museum items ported into the Unreal Game engine will be also shown.

Leave a Reply