Objective
The project aims at the development of theoretical foundations, specialized mathematical apparatus and algorithmic implementation of the information technology for automation and standardization of medical image analysis. Different approaches to medical image analysis will be conveniently integrated in order to obtain a common, more general and powerful, procedures for image analysis. As an illustration of efficiency of the developed theoretical methods, analysis of microscopic images of blood cells and lymphatic tissues will be considered. The objective analysis of biomedical (cytological and histological) images has been the subject of research for many years. The importance of this research activity can be easily appreciated from the example of Belarus, where the problem of diagnosis of people exposed to radiation is especially urgent and where about 3000 people are being operated annually. The case in point is hematology and oncology disease diagnostics. Biomedical image analysis here is necessary for extracting essential information ensuring the correct diagnosis. One of the most difficult tasks in biomedical image analysis is automated object extraction from the images and subsequent classification. The critical problem for automation of biomedical image analysis is to create and investigate algebraic structures for representing image models as objects of analysis and recognition, and image transform models as tools for efficient synthesis and implementation of basic procedures of image processing and analysis.
The practical goal of research is to develop:
· a prototype system or a toolkit to solve main tasks of automated color medical image analysis using novel image algebras techniques, efficient algorithms for image matching, segmentation approaches, neural networks, and automatic raster-to-vector transformation;
· a new information technology for the morphological analysis of the lymphoid cell nuclei of patients with hemopoietic tumors based on the combined use of pattern recognition and image analysis techniques.
In order to assess and verify the approaches and methods developed, a medical image database, provided by RWTH, will be used. This database contains currently over 12,000 images, which are classified according to their modality, body part examined and biological system. Furthermore, morph metric assessment of biological objects by quantitative parameters characterizing the changes of cell nuclei, fibers and tissues will be included to support the formulation of decision rules for diagnostics. The algorithms will be characterized by moderate computational complexity and allow the user to perform quantitative object analysis, which will have a noticeable impact. The project includes both theoretical and practical aspects of research carried out by the partners who have long-term experience in the different areas of medical image processing and analysis.
Call for proposal
Data not availableFunding Scheme
Data not availableCoordinator
56124 Pisa
Italy