+7 (495) 987 43 74 ext. 3304
Join us -              
Рус   |   Eng


Filimonova E.

Cand. Sci. (Eng.), Associate Professor, Digital Economy Department, Synergy University
Moscow, Russia

Approach to enhance the sharpness of the image by separating the contour information of software and hardware

The article is devoted to the current problem of sharpening the reproduced image. The article describes the processing of image information with the aim of sharpening. To increase the sharpness of the image, it is proposed to use the method of selection of the contour information of the original with its subsequent preservation and combination with the original image. To highlight the contour information offers the choice of the optimal operator for image segmentation. Image segmentation using information extraction operators was carried out in the MATLAB software environment. The work investigated the quality of the Sobel operator, the Prewitt operator, and the laplacian gaussian, since other operators do not meet the necessary requirements for the accuracy of contour reproduction. To assess the quality of work and accuracy of reproduction by the operators of part contours, a test object was created and used, which was created using the Adobe Illustrator software environment. The article presents the main stages of the proposed technique: image discretization, selection of contour information, combination of contour information with the original image. As a result of the work carried out, the criteria for obtaining contour information were determined, at which image sharpness would be enhanced: high resolution of the original source; lack of coarse discretization of contour lines; precise alignment of the contour information with the original image. For the selection of contours, the Sobel operator was chosen as the most accurately selecting contour. For the selected images, the proposed method is applied, prints are obtained. Prints are evaluated by a group of experts. Expert evaluation of impressions shows that the use of methods to improve the sharpness of the image gives unambiguously positive results.

Study of moire formation in the reproduction of information using software methods

The article is devoted to the current problem of the study of moiré formation in the reproduction of information using software methods. As a result of the study, the structures that cause the greatest appearance of moiré when rastering the image of the test object were identified. Also, the main regularities were revealed: strong noise of raster structures leads to loss of the image of the test object; the spectra of fields with noticeable moiré formation are modified due to the appearance of extra harmonics, independent of the semantics of the image; the more clearly and noticeably moiré in the image, the more the spectrum view changes, in which additional spectral lines appear; the smaller the difference between the optical density of the background and the optical density of the stroke, the more structural transformations caused by the interaction of the lattice and raster structure. On the basis of the study, the following conclusions were formulated: the use of an irregular raster structure leads to the absence of moiré in the image, blurring on the spectrum; the appearance of moiré when rastering the test object with stochastic structures indicates the presence of a regular component in them. In addition, the field was identified, while maintaining the bar structure, periodicity, lionachescu on the spectrum.

Investigation of color differences in the reproduction of memorable colors on visualization devices

This work is devoted to the study of changes in color coordinates on various visualization devices during color reproduction, in particular smartphones, as one of the most commonly used devices in the modern world, which is associated with the hardware dependence of the color reproduction system. The purpose of the work is to select visualization tools, determine their technical characteristics, determine test colors for visualization on various devices, determine the tolerances in reproduction of each color when using various viewing devices. To achieve the goal, such tasks were set as the selection of images containing memorable colors. These colors are fundamental in determining the tolerances in color reproduction, since information about them is inherent in each person on the basis of his life experience and knowledge, and a change in the reproduction of memorable colors, a violation in color rendering, is the most visually noticeable. Memorable colors are converted into samples – test objects, which are used to determine coordinates in a device-independent color space. Determination of tolerances is made when changing color reproduction through the use of selective (color) correction. When solving the problem, it was noted that there are colors in which, with small changes in color coordinates, the visual component changes to a sufficiently strong degree, while other colors, with a numerically identical change, do not visually change. For the selected colors on various visualization tools, the difference in color reproduction is calculated and the calculations of color differences for different models of viewing devices with a visual difference in perception are given. Read more...

Building and analyzing a machine learning model for short-term bitcoin market forecasting based on recurrent neural networks

In this article, the construction and analysis of machine learning models were performed for short-term forecasting in the cryptocurrency market on the example of bitcoin – one of the most popular cryptocurrencies in the world. The initial data for the study leads to the conclusion that over the long period of its existence, bitcoin has shown a high degree of volatility, especially evident in comparison with traditional financial instruments. The article substantiates that this market is influenced by a multitude of factors. No one can say for sure what makes up the value of a particular cryptocurrency, as it involves a range of reasons, which cannot be fully taken into account. To overcome this problem, we have considered the principle of recurrent neural network. It is described why networks with memory are better at making predictions on the time series than conventional autoregressive model and standard forward propagation networks. The initial data processing algorithm and transformation methods are defined. The sample was reduced in order to increase the speed of the network, by reducing the number of recalculations of weights. The algorithm of the family of recurrent neural networks was built and trained to test the hypothesis about their better adaptivity due to short-term and long-term memory. The model is evaluated on the test data representing the bitcoin exchange rate for 2021–2022, since this period is characterized by high volatility. It is concluded that it is reasonable to use a similar type of models for short-term forecasting of cryptocurrency rates. Read more...