Z.О. Solovieva, М.А. Skedina, V.K. Ilyin, N.V. Verdenskaya, I.A. Ivanova
The problem of recognition of an image of microbic cells in system of the automatic analysis is ra-ther actual. It’s solution opens opportunities to use method of visual recognition of microbial im-ages in order to obtain operative information on human microflora.
Specificity of bacterial cells, in comparison with other biological objects investigated by means of light microscopy, concerns, first of all, the small size of these objects. Other peculiarity of investigated images is low concentration of them in sample. Besides, images can form conglom-erates in microscope field.
Program opportunities of modern systems of computer processing of images of microobjects are great enough. However, considering specificity of analyzed images of preparations, it is necessary to revise requirements to the following items:
automatic focusing of a microscope,
the software of management of a microscope,
algorhythms of objects detection,
algorhythms of segmentation of the analyzed matter, both concerning quality, and velocity of treatment.
As a result of the study the software is designed, which allows to obtain authentic data at number of bacterial cells differing on morphological and tinctorial features. Such quality is im-portant to validate infection thread in time.
For increase of accuracy of the given method the further improvement of software is required.