Projet de recherche doctoral numero :3278


Date depot: 1 janvier 1900
Titre: 3D and semantic web: new tools to document artifacts and to explore cultural heritage collections
Directeur de thèse: Matthieu CORD (ISIR (EDITE))
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Non defini

Resumé: This thesis will investigate a methodology for the acquisition and processing in order to document heterogeneous objects, in particular works of art. The research will involve: 1. Research in acquisition, documenting, searching technologies (text and content) and semantic browsing of collections. This work will take into account ongoing research in current national and European projects; in particular 3D-Coform and EROS 3D. 3D-CoForm is a FP7 EU funded project that aim to advance the state of the art in 3D digitisation and to make the 3D an everyday practical choice for the documentation in the cultural heritage sector. EROS3D is an ANR French funded project working on the search by content using 3D models for cultural heritage. The EROS database at the Centre de Recherche et de Restauration des Musees de France is a good example of cultural heritage collection. It contain different media and different format and could represent a goodexample of heterogeneous database. 2. Integrating current technologies into practical business aspects: work-flows, procedures, planning of executions and business models for the exploitation of the 3D assets. From the point of view of data organization, the challenge is the investigation of new ways to represent and organize metadata, experimenting already existing tools like “Fedora” in combination with different ontologies as Dublin Core, CIDOC-crm, KOS, etc. From the point of view of 3D data acquisition, modelling and processing the definition of protocols and good practices applied to the problematic in this domain is at its early stages. 3. Proposing a methodology that cover all the aspects of the documentation pipeline: -- 3D acquisition: which technique is better in which case, why use laser, time of fly or photogrammetry in the different cases, which illumination for the acquisition, data registration and fusion, how to use multimodal acquisitions (multi source 3D acquisition, high definition images, etc); --3D modeling: best practices in data representation, best quality vs best performances, simplification, 3D watermarking; -- documenting: organisation of data, 3D tagging, semantic segmentation; -- exploitation of the data: how to present results, allow transparent access to data for end user, allow in deep access to data for scientific research, use of 3D model for conservation purposes, case studies and problematic; -- search services: data comparison, search by similarity, ranking, semantic classification.

Doctorant.e: Pitzalis Denis