Wall-facing Automatic images identification Laboratory
A quantitative analyses method for the study of ancient architecture
Acronym: W.A.L.(L)
Description
The W.A.L.(L) project aims to develop quantitative analysis methods and applications of Machine Learning to ancient architecture. Case study for the experiment and the development of a methodology are some Cretan wall structures mainly from the Iron Age (10th - 8th century BC).
It is an international project co-directed by the CNR ISPC and the Université Catholique de Louvain. The research involved numerous partners active in excavation and research missions at archaeological sites in Crete, such as the University of Catania, the Anavlochos Project (École française d'Athènes) and the Ghent University, as well as the CNR IIT. In particular, the wall samples on which the project focused fall on the sites of Phaistos, Haghia Triada, Sissi and Anavlochos.
The method developed by the W.A.L.(L) project, which can be replicated for architectures of other contexts and chronologies, contemplates several steps:
- the 3D photogrammetric survey of the walls,
- the manual segmentation of the virtual models and the classification of the different components in the Blender environment,
- the extraction of numerical features of the building materials through an add-on purposely created,
- the export of these values in a CSV format,
- the creation of a relational database,
- the creation of an interface by using the Django framework, aimed at the database population and data queries.
Machine Learning algorithms have been tested for the automatic recognition of some classes of building materials.
More info
Project Wall facing Automatic Image Identification Laboratory – W.A.L.(L)by Francesca Buscemi, Giovanni Gallo, Marianna Figuera, Yaser Gholizade Atani, Angelica Lo Duca, presented at the GCH 2023 – Eurographics Workshop on Graphics and Cultural Heritage
Project information
- Italian project leader: CNR ISPC
- Foreign project leader: Université Catholique de Louvain
- ISPC Scientific coordinator: Francesca Buscemi
- Partners: Università di Catania, Anavlochos Project (École française d’Athènes); Ghent University
- Involved Countries: Italia – Grecia
- Site/Area/Case study: Creta sites: Festòs, Haghia Triada, Sissi, Anavlochos
- Status: ongoing
- Funding: CNR Joint Lab (2021 – extended until 2023)
