Quality of word and concept embeddings in targetted biomedical domains.

Embeddings are fundamental resources often reused for building intelligent systems in the biomedical context. As a result, evaluating the quality of previously trained embeddings and ensuring they cover the desired information is critical for the success of applications. This paper proposes a new evaluation methodology to test the coverage of embeddings against a targetted domain of interest.

Semantic annotation of 3D anatomical models to support diagnosis and follow-up analysis of musculoskeletal pathologies.

Purpose: While 3D patient-specific digital models are currently available, thanks to advanced medical acquisition devices, there is still a long way to go before these models can be used in clinical practice. The goal of this paper is to demonstrate how 3D patient-specific models of anatomical parts can be analysed and documented accurately with morphological information extracted automatically from the data. Part-based semantic annotation of 3D anatomical models is discussed as a basic approach for sharing and reusing knowledge among clinicians for next-generation CAD-assisted diagnosis and treatments.

Between algorithm and model: different Molecular Surface definitions for the Poisson-Boltzmann based electrostatic characterization of biomolecules in solution.

The definition of a molecular surface which is physically sound and computationally efficient is a very interesting and long standing problem in the implicit solvent continuum modeling of biomolecular systems as well as in the molecular graphics field. In this work, two molecular surfaces are evaluated with respect to their suitability for electrostatic computation as alternatives to the widely used Connolly-Richards surface: the surface, an implicit Gaussian atom centered surface, and the surface. As figures of merit, we considered surface differentiability and surface area continuity with respect to atom positions, and the agreement with explicit solvent simulations.