M.V. Dyudin - Post-graduate Student, Department of Biomedical Engineering, Southwest State University (Kursk) S.A. Filist - Dr. Sc. (Eng.), Professor, Department of Biomedical Engineering, Southwest State University (Kursk). E-mail: SFilist@gmail.com P.S. Kudryavtsev - Post-graduate Student, Department of Biomedical Engineering, Southwest State University (Kursk). E-mail: 79pavel97@mail.ru

In article intellectual systems of the analysis flyuorogramm a thorax which principle of work is based on a hybrid campaign are consi-dered. For allocation of contours of lungs the way based on «sliding» two threshold a binarization of the initial image of a x-ray picture and the morphological analysis was offered. The essence of the «sliding» two threshold binarization consists in division of range of brightness of pixels 0-255 on subranges with formation of «two threshold windows» between top and lower thresholds. From a set of N bitmaps the bitmap which is most adequately describing pixels, the corresponding areas on the flyuorogramma occupied by lungs is allocated. Process of search of this image carries out the two-alternative qualifier constructed on the basis of correlation methods of identification. Then morphological processing of the image which allows to remove objects to which the objects adjoining an image edge and objects which sizes there is much less than sizes of lungs belong from the image «superfluous» is carried out. Using the morphological operator imfill (fills openings on an initial bitmap), we receive the negative image of the areas adjoining an image edge. Having increased a grayscale image by a bitmap, we receive the image with the remote fragments adjoining an image edge. After obtaining the one-pixel image of lungs start the spectral analysis of contours of the left and right lung. For carrying out the pla-nimetric analysis of RS of lungs Fourier\'s descriptors on the basis of which the space of informative signs intended for neural network classification of radiological syndromes was formed were used.

 

1. Medicinskaja rentgenologija: tekhnicheskie aspekty, klinicheskie materialy, radiacionnaja bezopasnost / Pod red. R.V. Stavickogo. M.: MNPI. 2003. 344 s.
2. Filist S.A., Tomakova R.A., Gorbatenko S.A. i dr. Analiz gistologicheskikh izobrazhenijj posredstvom morfologicheskikh operatorov, sintezirovannykh na osnove preobrazovanija Fure i nejjrosetevogo modelirovanija // Biotekhnosfera. 2010. № 3(9). S. 54−60.
3. Tomakova R.A., Filist S.A., Rudenko V.V. Nechetkaja setevaja model intellektualnogo morfologicheskogo operatora dlja formirovanija granic segmentov // Nauchnye vedomosti BelGU. Belgorod: Izd-vo BelGU. 2011. № 1(96). Vyp. 17/1. S. 188−195.
4. Tomakova R.A., Efremov M.A., Filist S.A.i dr. Gibridnye tekhnologii vydelenija medlennykh voln iz kvaziperiodicheskikh signalov // Izvestija JUgo-Zapadnogo gosudarstvennogo universiteta. 2011. № 1(34). S. 66−73.
5. Ivanov A.V., Tomakova R.A., Mishustin V.N. Intellektualnaja sistema morfologicheskogo analiza patologicheskikh izmenenijj pri retinopatii // Izvestija JUgo-Zapadnogo gosudarstvennogo universiteta. 2012. № 2(41). CH. 1. S. 19−27.