The objective of AIM@SHAPE is to build a shared conceptualisation of a multi-layered architecture for shape models, where the simple geometry is organized in different levels of increasing abstraction: geometric, structural and semantic layers. The key actions to reach this objective are :
- the harmonization of the approaches to shape modelling in the reference communities (i.e., Computer Graphics and Computer Vision) via the definition of shared vocabularies and ontologies for reasoning and processing shape data;
- the adoption of a research plan focused on studying methods to preserve semantic content, to annotate automatically digital shapes, to interact with the semantics, and to maintain and update the semantics at the different stages of digital shape lifecycles.
The research activities of AIM@SHAPE are addressing a number of open research problems that occur in the two main lifecycles of digital shapes: the first goes bottom-up from the acquisition of 3D objects up to the semantic level, while the second corresponds to the top-down design from a concept defining an object at the semantic level down to the digital model. AIM@SHAPE ambition is to provide multi-dimensional content in an active, or even proactive, sense by revisiting the whole lifecycle of digital shapes with emphasis on the acquisition and modelling pipeline: it is easier to convey and share shape knowledge, if knowledge is preserved and modelled with the digital shape. The research plan of AIM@SHAPE is focused on studying methods to preserve semantic content, to annotate automatically digital shapes, to interact with the semantics, and to maintain and update the semantics at the different stages of digital shape lifecycles. The research activities are grouped into three programmes dealing with the acquisition and reconstruction, the analysis and structuring, and the interpretation and mapping of shape models in given domains of knowledge.
A purely geometry-based representation of a digital shape can be used to view the shape (a); a structural view can give hints and show how the shape components are linked together (b); a semantic view is able to propose an interpretation or meaning of the digital shape (c).
The research activities address the development of methods and tools for switching from one level to another, trying to preserve, extract and code shape knowledge during acquisition and reconstruction processes, analysis and structuring processes, interpretation and mapping processes.
Acquisition and Reconstruction
Shapes may enter the digital world either by digitalisation or by design processes. In both cases, there might be a considerable amount of knowledge available about the shape, which is neither captured nor coded, in the digital representation of the shape. The interest here is on the development of efficient mechanisms for gathering shape-related knowledge at the very beginning of the shape lifecycle.
Analysis and Structuring
The structure of a shape is obtained by organising the geometric information and by making explicit the association between parts or components of shape models or shape data. If the organisation is geometry-based it is possible to cite as examples: multi-resolution models, multi-scale models, curvature based surface decompositions, topological decompositions, etc. If the approach is already semantic oriented, it ius possible to list: shape segmentations, pattern or cluster based structuring, form feature representations, etc. Research topics belonging to this cluster cover issues related to preserving and enhancing shape information during geometric processing, and to effectively capturing the structure of a shape by identifying relevant shape components and their mutual relationships.
Interpretation and Mapping
At the semantic level there is the association of a specific meaning to structured and geometric models through annotation of shapes, or shape parts, according to the concepts formalised by the domain ontology. For example, in the manufacturing domain, the association of the semantics to a product model is done through the detection of all the shape features which have a specific definition in the manufacturing ontology (e.g., slots, steps or notches). Therefore, a semantic model is the representation of a shape embedded into a specific context. Research topics belonging to this cluster cover issues related to the automatic annotation of digital shapes, to mapping semantic descriptions of shapes from one domain of knowledge to another, to maintaining the semantics of existing shape models
Ontologies and Metadata for Shape Models and Tools
The core of the integration pursued by AIM@SHAPE resides in the homogenisation of the approach to modelling shapes and their associated semantics using knowledge formalisation mechanisms, in particular metadata, which describe shape models and shape processing tools, and ontologies, which provide the rules for linking semantics to shape or shape parts. Through a common formalization framework, it will be possible to build a shared conceptualisation of a multi-layered architecture for shape models, where the simple geometry is organized in different levels of increasing abstraction: geometric, structural and semantic layers.
To get details about the mobility exchange reports click here