Quantitative Analysis of Structural Changes in Steel Caused by High-Temperature Holding in Hydrogen

We give a quantitative estimate of changes in the structure of 15Kh2MFA steel after its long-term exposure to the combined action of high temperature, tensile stresses, and hydrogen as the factors of degradation of this steel under conditions of oil hydrocracking. Under laboratory conditions, we study the effect of high-temperature hydrogen degradation of 15Kh2MFA heat-resistant steel on changes in the quantitative characteristics of its structural elements, in particular, in the dimension and average distance between carbides as well as in their orientation from one grain to another. Algorithms for the solution of the formulated problems are described, and the possibilities of proposed approaches for the quantitative automatic processing of metallographic images are shown.     

Automated system for the analysis of circular spectra of metallographic images

We propose a new approach to the determination of the orientation of carbides in steel based on the method of circular spectrum. The software aimed at the analysis of metallographic images of 15Kh2MFA steel on the basis of the circular spectrum of the image and its principal characteristics (variance and derivative) is developed. By using this software, it is possible to determine the orientations of carbides from grain to grain.Translated from Fizyko-Khimichna Mekhanika Materialiv, Vol. 40, No. 2, pp. 113–115, March–April, 2004.     

Features in the design of optimal recognition systems

Features that are inherent to recognition systems that lay claim to constituting an optimal design are considered. A maximal probability of correct recognition is characteristic of such systems. The degree of optimality of systems, the quality of a constructed classifier, or the dimension of the confidence interval may be established on the basis of these features.     

Evaluation of the characteristics of cracks according to the images of the surfaces of materials

We consider the problem of three-dimensional reconstruction of the surface according to its two-dimensional images. The results of this reconstruction are applied to the analysis of the images of materials with cracks. We describe each step of the algorithm used for the analysis of the characteristics of cracks in the material according to their images. It is shown that the data on the profile and volume of the cracks can be obtained with the help of the proposed methods. __________ Translated from Fizyko-Khimichna Mekhanika Materialiv, Vol. 43, No. 4, pp. 107–111, July–August, 2007.     

Changes in the central hemodynamics in the surgical treatment of abdominal aorta aneurysm

The algorythm of examination and treatment of patients with an abdominal aorta aneurysm was proposed.     

The ultrasonographic confirmation of the role of external carotid artery in blood supply of ischemic lesion in the big cerebral hemisphere

Using intraoperative monitoring of the blood flow linear speed in a. cerebralis media while total a. carotis interna occlusion in 10 patients the authors have concluded, that the degree of blood flow compensation in a. cerebralis media depends not only on collateral compensation possibilities by the contralateral side and rate of it inclusion, but largely-on a. carotis externa possibility on the lesion side.     

Tomographic Image Reconstruction Based on Generalized Projections

The method of tomographic reconstruction of inhomogeneities in case of an arbitrary directional diagram and scanning along one coordinate is described. It is shown that in this case registered data (projections) are represented as the sum of the line-by-line convolutions of the lines of the directional diagram and the corresponding line of the reconstructed distribution. The features of projection data, backprojections are analyzed. It is proposed to construct a cumulative image as an additive or conjunctive combination of the backprojection results for different orientation diagrams. An iterative procedure for constructing sequential approximations to the desired solution is proposed.

     

Upper-bound estimates for classifiers based on a dissimilarity function

Approaches to calculating upper-bound estimates of recognition probability are proposed that can be used for a more general class of models. One of estimates determines the stability of object coverage by classification algorithms on the basis of distribution of distances between objects, and another estimate is underlain by leave-one-out cross-validation. This considerably simplifies and facilitates the construction of estimates.