Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS

Тел.: +7 (495) 625-9241


Automation of medical images processing for diagnose of pathology


О.R. Nikitin, А.S. Pasechnik, Salem Al Saidi

Abundantly clearly that any disease is simpler to treat at its earliest stages, in this case development of complications in pathology less possibly, and process of treatment proceeds quicker. Diagnosing of a various type of diseases in this case should be as much as possible operative, without distracting a significant amount of time and means of the patient and consequently, it should have wide mass character. The increase in loading at the medical diagnostic centers leads to decrease in their capacity that negatively affects, including on the social relation of patients to diagnosing procedure. The situation can be improved considerably, everywhere introducing complexes and systems for automation of processes of service of patients in the medical centers. Differentiation at detection of various types of objects on the studied image can be reached by a way of installation of the corresponding managing directors of parameters of methods of contouring and segmentation. To such parameters of a preliminary filtration of the video image, threshold levels of sensitivity of contouring algorithm corresponding to concrete type of studied objects, and also a number of technological parameters of algorithm of the segmentation, causing its fine tuning belong to a concrete diagnostic situation, first of all, contouring method, the most optimum for this type of the video image. Formation of library of operating parameters of algorithms of contouring and segmentation possible in two ways. One of them is that the main part of library is created before operation of the automated program complex by test researches, with involvement of profile experts in the field of diagnostics. For this purpose it is necessary to designate previously entry diagnostic conditions, to define quantity of possible types for diagnostics of the abnormal objects having basic differences among themselves on the morphological properties and structure. The second way of formation of base of operating parameters of algorithms of contouring and segmentation is that this library is open and can be added already directly in use a complex. One of the most important tasks of system of the automated detection of abnormal objects on video images, is their localization. Localization is understood as procedures of determination of coordinates of border of studied object on the image, calculations of its geometrical sizes, and also a relative positioning of objects on the video image. Result of segmentation of abnormal object on the image is the binary matrix of the dimension corresponding to dimension of the image, having elements, distinct from zero, in those places where on the image there is a border of the allocated object. Uniting the reached results at research of possibility of automation of procedures of contouring, segmentation and localization, it is possible to offer the structural organization of a hardware-software complex of the automated processing of images for diagnostics of pathologies. Such complex can't independently be used without participation of the doctor – the expert as nevertheless there is a risk of pronouncement of the incorrect diagnosis. Diagnostic data arrive from studied bio object by means of the device of collecting primary information. Introduction of similar complexes and their integration with the existing hardware base which is available in the medical diagnostic centers, will allow to increase essentially quality of medical care in such establishments, to reduce time of carrying out procedure of diagnosing for 40 %, and thus to make high-precision operative diagnostics of pathologies more available to a general population.

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio