Image processing methods and architectures in diagnostic pathology.

Abstract

Grid technology has enabled the clustering and the efficient and secure access to and interaction among a wide variety of geographically distributed resources such as: supercomputers, storage systems, data sources, instruments and special devices and services. Their main applications include large-scale computational and data intensive problems in science and engineering. General grid structures and methodologies for both software and hardware in image analysis for virtual tissue-based diagnosis has been considered in this paper. This methods are focus on the user level middleware. The article describes the distributed programming system developed by the authors for virtual slide analysis in diagnostic pathology. The system supports different image analysis operations commonly done in anatomical pathology and it takes into account secured aspects and specialized infrastructures with high level services designed to meet application requirements. Grids are likely to have a deep impact on health related applications, and therefore they seem to be suitable for tissue-based diagnosis too. The implemented system is a joint application that mixes both Web and Grid Service Architecture around a distributed architecture for image processing. It has shown to be a successful solution to analyze a big and heterogeneous group of histological images under architecture of massively parallel processors using message passing and non-shared memory.