Computer Optics

853-863

Authors: A.V. Volyar, E.G. Abramochkin, M.V. Bretsko, Ya.E. Akimova, Yu.A. Egorov

In general, a standard Laguerre–Gauss (LG) beam, whose state is given by two quantum numbers (n, ????): the radial number n and the azimuthal number ???? (or the topological charge (TС) of the vortex carried by the LG beam), is unstable with respect to weak perturbations. This is not difficult to see if we decompose the complex amplitude of the LG beam in terms of Hermite–Gauss modes (HG), with the total number of HG modes being equal to N = 2n +???? + 1. If we now slightly change the amplitudes and phases of each HG mode, then the structure of the LG beam radically changes. Such a combination of modes is called a structured LG beam (sLG), which can carry large additional arrays of information embedded in the sLG beam by encoding the amplitudes and phases of the HG modes (excitation of modes). But as soon as a perturbation is inserted into the LG beam, its orbital angular momentum (OAM) can change dramatically in such a way that the value of the OAM changes in the interval (–????, ????), and the total TC – in the interval (–2n – ????, 2n + ????). At n = 0, the OAM changes smoothly in the interval (–????, ????), however it is worth "turning on" the radial number n, as the OAM oscillations occur. The number of minima (maxima) of the oscillations is equal to the radial number n in the interval θ = (0, π) and θ = (π, 2π), with their amplitude nonlinearly depending on the difference ???? – n, except for the point θ = π, where the structured beam becomes degenerate. If ???? = 0, then the OAM is zero, so that in the sLG beam structure, we observe either a symmetrical array of vortices with opposite-sign TCs or a pattern of edge dislocations, the number of which is equal to the radial number n. We also found that, despite the fast oscillations of the OAM, the absolute value of the total TC of the sLG beam does not change with variation of both the amplitude ε and phase θ parameters, but depends solely on the initial state (n, ????) of the LG beam and modulo (2n + ????).

872-876

Authors: V.V. Kotlyar, E.G. Abramochkin, A.A. Kovalev, A.A. Savelyeva

We show here that the product of two Laguerre-Gaussian (LG) beams, i.e. double LG beams (dLG), can be represented as finite superposition of conventional LG beams with certain coeffi-cients that are expressed via zero-argument Jacobi polynomials. This allows obtaining an explicit expression for the complex amplitude of the dLG beams in the Fresnel diffraction zone. Generally, such beams do not retain their structure, changing shape upon free-space propagation. However, if both LG beams are of the same order, we obtain a special case of a "squared" LG beam, which is Fourier-invariant. Another special case of the dLG beams is obtained when the azimuthal indices of the Laguerre polynomials are equal to n – m and n + m. For such a beam, an explicit expression is obtained for the complex amplitude in the Fourier plane. We show that if the lower indices of the constituent LG beams are the same, such a double LG beam is also Fourier-invariant. Similar to conventional LG beams, the product of LG beams can be used for optical data transmission, since they are characterized by azimuthal orthogonality and carry an orbital angular momentum equal to the topological charge.

884-891

Authors: R. Ilinsky

A ball lens is analyzed. The refractive index of this lens is a function of the distance from the ball center. Conditions are derived under which such a lens is free of the third-order spherical aberration given a beam of incident parallel rays.

899-904

Authors: M.M. Hamza, V.A. Blank, V.V. Podlipnov, L.L. Doskolovich, R.V. Skidanov, B. Fan

A device for visualization of blood vessels in the human skin is presented. A diffractive optical element is used to locate blood vessels in the skin image. It has been shown that visualization of blood vessels is most effective in the form of obtaining an index image at wavelengths of 735 nm and 835 nm. The index is calculated using a formula similar to the NDVI formula. The work also uses an application software that is used to solve the problems of spectral analysis.

914-920

Authors: V. Pavlov, A. Tuzova, A. Belov, Y. Matveev

Many different filters can be used to reduce multiplicative speckle noise on radar images. Most of these filters have some parameters whose values influence the result of filtering. Finding optimal values of such parameters may be a non-trivial task. In this paper, a formal automated method for finding optimal parameters of speckle noise reduction filters is proposed. Using a specially designed test image, optimal parameters for the most commonly used filters were found using several image quality assessment metrics, including the Structural Similarity Index (SSIM) and Gradient Magnitude Similarity Deviation (GMSD). The use of filters with optimal parameters allows processing (detection, segmentation, etc.) of radar images with minimal influence of speckle noise.

929-938

Authors: S.V. Sai, A.V. Zinkevich, E.S. Fomina

The article describes features of digital processing of image signals in the process of coding based on discrete cosine (DCT) and wavelet transforms (DWT) that are used in the JPEG and JPEG2000 compression standards. To compare DCT and DWT, a digital model of the system has been developed that implements the same stages of signal processing, except for the stages of the proper discrete transforms. A method for analyzing the efficiency of the transformations based on objective assessments of image quality depending on the compression ratio is proposed. The peculiarities include the fact that, in contrast to the popular PSNR and SSIM metrics, it is proposed that the quality be assessed using the reduction factor for the RAW image format, the calculation of which is associated with the contrast sensitivity of vision. As a result of the research, quantitative estimates of the compression efficiency are obtained for the given quality parameters, depending on the type of conversion and the detail of the RAW images. Recommendations are made regarding the development of methods for increasing the efficiency of image compression based on DWT or DCT.

948-954

Authors: V.V. Kapustin, A.S. Zahlebin, A.K. Movchan, M.I. Kuryachiy, M.V. Krutikov

This paper considers an experimental study of the layout of an active-pulse television measuring system in the problem of assessing the accuracy of measuring the distance to objects using the depth maps. The main technical characteristics and structure of the active-pulse television measuring system layout are described, the description of the multi-zone ranging method used in the experiment is given. The field tests were carried out using a system for terrain orthophotomaps construction by an unmanned aerial vehicle and a geodetic measuring instrument, which is a reference for building a terrain plan and fixing distances between objects on the ground. The technique of carrying out aerial work is described to obtain the necessary data array, on which a digital model and an orthophotomap of the area were subsequently built. Conclusions are drawn about the accuracy of digital terrain models built based on the results of aerial photography from an unmanned aerial vehicle with a geodetic receiver on board and the applicability of these data as reference data for testing a prototype of an active-pulse television measuring system.

963-970

Authors: E.Yu. Shchetinin

Early detection of patients with COVID-19 coronavirus infection is essential in ensuring an adequate treatment and reducing the burden on the health care system. An effective method of detecting COVID-19 is computer analysis of chest X-rays. The paper proposes a methodology that consists of stages of formatting X-ray images to the size (224, 224) size, their classification using deep convolutional neural networks, such as Xception, InceptionResnetV2, MobileNetV2, DenseNet121, ResNet50 and VGG16, which are pre-trained on the ImageNet dataset and then fine-tuned on a set of chest X-rays. The results of computer experiments showed that the VGG16 model with fine-tuning of parameters demonstrated the best performance in the COVID-19 classification with accuracy = 99.09 %, recall = 99.483 %, precision = 99.08 % and f1_score = 99.281 %.

980-987

Authors: E.Y. Shchetinin, A.G. Glushkova

Due to cardiovascular diseases millions of people die around the world. One way to detect abnormality in the heart condition is with the help of electrocardiogram signal (ECG) analysis. This paper's goal is to use machine learning and deep learning methods such as Support Vector Machines (SVM), Random Forests, Light Gradient Boosting Machine (LightGBM), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BLSTM) to classify arrhythmias, where particular interest represent the rare cases of disease. In order to deal with the problem of imbalance in the dataset we used resampling methods such as SMOTE Tomek-Links and SMOTE ENN to improve the representation ration of the minority classes. Although the machine learning models did not improve a lot when trained on the resampled dataset, the deep learning models showed more impressive results. In particular, LSTM model fitted on dataset resampled using SMOTE ENN method provides the most optimal precision-recall trade-off for the minority classes Supraventricular beat and Fusion of ventricular and normal beat, with recall of 83 % and 88 % and precision of 74 % and 66 % for the two classes respectively, whereas the macro-weighted recall is 92 % and precision is 82 % .