The problem of calculating a quantitative quality assessment of digital images of metal object fractures recorded by a camera or a digital microscope is considered. Quality assessment is performed when the reference images are absent. The paper presents an approach based on calculation of local estimates followed by analysis of their distribution. Several variants for calculation of local estimates have been studied. Those whose distribution is unimodal were selected. It is shown that the average of local estimates is an acceptable general characteristic of image quality if they have a normal (Gaussian) distribution. In this case, the average is one of its parameters. Otherwise, the parameters of the Weibull distribution can serve as more accurate quantitative characteristics of image quality in general. The proposed approach divides more objectively the set of the analyzed images into two groups those with satisfactory or unsatisfactory quality for performing expert studies using images. Examples of the quality assessment of different object images recorded at different resolutions are presented.

For people re-identification in distributed video surveillance systems, it is important to have an algorithm that ensures efficiency when a person is blocked by other people or objects. Therefore, for this task, an algorithm has been developed that involves the compound descriptor formation, which includes a global features vector of a person’s image and three local ones for its upper, middle and down parts. The regions of interest selection is carried out based on the detecting key points results in the image of the human body. If a person's image part is occluded by other people or objects, then it is classified as invisible. The hidden part person image is not used in the formation of the local feature. To obtain it, the average value of the k-nearest neighbors such features of a person’s image is calculated. The experiments performed indicate an increase in re-identification accuracy for the Market-1501, DukeMTMC-ReID, MSMT17 and PolReID1077 datasets.

A mathematical model was developed, during which the voltage equations of the three-phase winding of a brushless DC motor with permanent magnets, the electromagnetic torque of a brushless DC motor, electromagnetic power, the induced emf of each winding, and the differential torque equation of the servo system were obtained. Based on the constructed mathematical model of the BLDC for a specific motor with given parameters, simulation modeling was carried out and the dependences of electromechanical quantities were obtained: angular velocity of the rotor, electromagnetic torque, stator phase current and rotor rotation angle on time. The simulation results confirmed the theoretical justification of the mathematical model. The resulting mathematical description of the BLDC can be used to build the architecture of a control unit based on a neural network controller.

When operating automated traffic control systems and their transformation into intelligent transport systems, modern requirements are imposed on the overall parameter safety, which characterizes the quality and efficiency of road traffic, especially when organizing high-speed loaded traffic of motor vehicles. The article discusses the use of various mathematical methods for these purposes when modeling transport processes and systems, including taking into account the development of artificial intelligence algorithms, so that when making decisions on traffic control, one can have reliable forecast indicators obtained from adequate models. A comparison of the models is made and recommendations are given on their applicability and the results obtained for traffic management purposes.

The article discusses the problem of studying the probabilistic characteristics of vibrations of the propellermotor group of an unmanned aerial vehicle such as a multicopter. The results of the research are presented in the form of quantitative values of the vibration parameters of the propeller-motor group of the agrodrone at given values of the input parameters of the electric drive control system, taking into account the influence of external factors. It is shown that the values of the parameters characterizing the vibrations of the propeller-motor group of the agrodrone can be different depending on the operating modes of the engines and changes in external conditions.

Based on the experimental studies carried out, sets of measured random vibration parameters were obtained, which made it possible, based on the application of mathematical statistics methods, to calculate the mathematical expectations and dispersions of the amplitudes and velocities of vibration processes. The obtained probabilistic characteristics make it possible to study emissions (crossing a given level) of processes, to evaluate the intensity and probabilistic characteristics of the process going beyond the given boundaries.

A mathematical model of the vibration process is substantiated, which allows, on the basis of the obtained experimental data, to assess the influence of vibrations on the elements of equipment installed on a multicopter, to assess its performance and reliability, and to determine tolerances on equipment parameters. The computer modeling of the probabilistic characteristics of vibration processes clearly demonstrated the efficiency of the proposed research methodology.

Modeling of the first principles of the electronic properties of complex oxides of rare earth elements (BaY₂O₄, BaGd₂O₄, BaLu₂O₄) as precursors of high-temperature superconductors has been performed. The VASP software package was used as a modeling environment, in particular the method of coupled plane waves (PAW method), which allows us to obtain fairly accurate results for calculating the electron density and band structure. From the analysis of the obtained band energy structure, it follows that the studied REE oxides have a band gap width Eg = 3.29–3.84 eV, which is characteristic for dielectric materials. The studied compounds based on these rare earth elements selected from the yttrium (Y, La, Gd–Lu) and cerium (Ce–Eu) groups are characterized by an increase in Fermi energy and a decrease in the band gap as the atomic number (39, 64, 71) of the element in the periodic table increases. A method for modeling the quantum layers of the studied materials by simulating the restriction of the crystal structure along one of the coordinate axes is proposed. This representation approximates the model of the crystal lattice of REE oxides to the situation of analyzing a quantum layer whose thickness is equal to the size of the crystal cell along the specified axis. The rupture of atomic bonds in a crystal is simulated by increasing the distance between atomic layers along this axis to values at which the value of free energy is stabilized. In the quantum layer of rare earth element oxide (with its thickness close to 1 nm), a wider range of energy values is formed in which electrons are distributed than is observed in the continuous version, and the expansion of the electron distribution area extends to the energy levels of the band gap. This is explained by the fact that the geometric discretization of nanoscale structures determines the discreteness of the quantum-dimensional energy spectrum.

The work discusses a number of techniques for segmenting dermoscopic images of skin lesions to identify the areas occupied by these lesions. Segmentation is necessary as the first stage of most methods of computer diagnostics of malignancy of neoplasms. A number of techniques, such as ABCDE, use the shape of the tumor as one of the criteria for making a diagnosis; for others, such as the use of convolutional neural networks, identifying the tumor allows one to increase the accuracy of the results obtained. The work discusses three methods of segmentation: thresholding using Otsu's method to calculate the threshold value, a convolutional neural network built on the U-net architecture, and a similar convolutional neural network with an added attention mechanism. The advantages and disadvantages of each method are considered, as well as the possibility of using them together to obtain the best segmentation results.

The paper considers the application of an algorithm based on a morphological projector for determining structural differences for comparing dermoscopic images. This will allow to identify changes that have occurred in skin lesions over time, for a more accurate diagnosis of their malignancy. The proposed algorithm makes it possible to detect differences in images even if there is a significant difference in the brightness and color levels of the compared images, and also ignores small insignificant details, such as noise, dermatoscope optics marks, hair, etc. A method for correcting the desynchronization of images using the structural similarity index as a similarity metric, and the sine-cosine algorithm as an optimization algorithm is proposed. The proposed algorithms were tested on dermatoscopic images and the possibility of their application was demonstrated.

The subject of research is the use of voice processing technology of the patient in IT medicine. The purpose of the article is to develop a neural network for the diagnosis of lung diseases using sound analysis of the patient's voice. The study includes training of a neural network, development of a mobile program for collecting patient sound, extraction of sound characteristics on the server side, diagnostics of sound data using a trained neural network and return of diagnostic results to the mobile application program. A block diagram of voice processing from the source signal to the extraction of an audio file is presented, as an example, the extraction of MFCC and FBank functions is given. The structure of a convolutional neural network (CNN), which was trained on a standard dataset of respiratory diseases, is given. A simplified process of classification of breathing sounds necessary for the prediction of lung diseases is given. For practical implementation, the VGGish network is used in the Python programming environment, which has network parameters trained using a data set. The experiments were carried out on the Android service framework platform, which is divided into two parts: Android front-end and server. The interface part implements the interactive user function and is responsible for entering audio data. After downloading the audio, the server will pre-process the audio, and call CNN to classify the audio, the results are returned to an external module on the smartphone. The total accuracy of the model reached 83.6 %.