Modelli Digitali 3D / Accessibilità e fruizione inclusiva | 3D Digital Models / Accessibility and Inclusive Fruition

The use of active/passive sensors allows surfaces to be discretized on the basis of point clouds associated with radiometric information. The validity of the available applications is widely documented: in the last two decades, scientific articles on the subject have grown from hundreds to tens of thousands. Polygon meshes generated directly (scan-to-BIM) or with the support of shared, interoperable processes can be navigated with powerful Real-Time Rendering engines to verify their characteristics within the image/video. Digital Twins (DT), when embedded within multimedia workflows and experiences (BIM-HBIM), can be managed by algorithms and mathematical models for the analysis and manipulation of geometric data, i.e., the modification of faithful and mirror copies of the digitally reconstructed product. The change in condition of the physical dual is retroactively controlled to (Field Information Modelling, FIM) record mutations, degradation or design changes in shapes. However, the massive availability of uninterpreted models requires an optimised development of methods for semantic segmentation and classification of data, information processing and analysis, as well as the integration of additional content from alternative sources and different sensor types. Ontologies of datasets - processed by Machine Learning (ML) and Deep Learning (DL) algorithms for the fast recognition of visual patterns (semantic segmentation) – have confirmed the validity of the services offered by the network: bringing the research of individuals into dialogue with Data Science is a necessity rather than an opportunity, especially for Cultural Heritage. The development of procedures that operate on the textures of models, projecting and visualising 2D results on the obtained geometries, also makes it possible to integrate and superimpose data and information from specialised investigations by experts in the field. The attention of research therefore focused on the evolution of digital language, and, contextually on the standardization of codes and processes to apply to web content. The latter should ensure widespread and dynamic access in compliance with European recommendations and guidelines. In this scenario, the proposed issue of DISEGNARECON intends to collect and compare contributions that, in the international panorama, illustrate the most recent scientific developments with respect to the treatment of polygon meshes and discrete models. Case studies will be opportunities to demonstrate the reliability of the achieved results, i.e., to discuss inconsistencies and limitations that can guide shared paths. For the reasons stated in the preamble, discussion topics will include: • investigations on the relationship between discrete and structured models to generate reversible types of networked and real-time rendered models for game-engines; • new museum exhibition and fruition techniques and multi-platform simulations (Common Data Environment) of experimental models for participatory dissemination; • methods for classifying 3D points useful for salient maps; • integration between survey and characterisation of 3D points derived from images/videos; • machine learning/deep learning applied to the processing of point clouds derived from image/video acquisition; • optimisation functions of algorithms from the computing environment inherent to mathematical modelling and neural network architecture for learning on graphs and pattern recognition for segmentation and semantic processing; • interactive circularity between acquisition, processing and inductive verification of processes; • development of standardised methodologies for morphological analysis, degradation mapping or enrichment and computerisation of data related to network services; • analysis and control of errors and/or inconsistencies (bias) in the input/output material of DL/ML/AI machines.