Semantic Integrative Digital Pathology: Microsemiological Semantics and Image Analysis Scalability

What is it about?

Being able to provide a traceable and dynamic second opinion has become an ethical priority for patients and healthcare professionals in modern computer-aided medicine. In this perspective, a semantic cognitive virtual microscopy approach has been initiated, by focusing on cognitive approaches for computational pathology. Besides ensuring the traceability of the second opinion provided and supporting the orchestration of high-content image analysis modules, the semantics will be crucial for the correlation between digital pathology and noninvasive medical imaging modalities. In addition, semantics has an important role in modeling the links between traditional microscopy and recent label-free technologies. The massive amount of visual data is challenging and represents a characteristic intrinsic to digital pathology. The design of an operational integrative microscopy framework needs to focus on scalable multiscale imaging formalism. In this sense, we prospectively consider some of the most recent scalable methodologies adapted to digital pathology as marked point processes for nuclear atypia and point-set mathematical morphology for architecture grading. To orchestrate this scalable framework, semantics-based WSI management (analysis, exploration, indexing, retrieval and report generation support) represents an important means towards approaches to integrating big data into biomedicine. This insight reflects our vision through an instantiation of essential bricks of this type of architecture. The generic approach introduced here is applicable to a number of challenges related to molecular imaging, high-content image management and, more generally, bioinformatics.

Featured Image

Why is it important?

Semantics will be the next very important dimension to consider after the actual tsunami in the area of deep learning. Its use will not only improve the traceability and the pathologists' acceptance but will also be able to give a context to the use of efficient deep learning tools, which lake of traceability and for which the transfer learning could thus be framed or guided.

Perspectives