Search

Walking is a good way to get to know a new district or city. However, it may be difficult for a tourist not familiar with the area to choose a route for a walk because they do not know, what places of interest there are nearby. This is especially true for areas unpopular with tourists and not mentioned in guidebooks. But even if attractions are listed there, it can be challenging to think of the best route to see them all. In this case, a special navigation service can help, which is able to build more interesting walking routes rather than shorter ones (as existing popular navigation services do). Such routes may be longer but they include local places of interest, green areas and embankments. The article proposes an algorithm that uses data on places of interest from Open Street Map and applies a modernized A* algorithm as well as a genetic algorithm to build such routes. Creation of two route options is considered: a route between two different points (if the tourist wants to reach a specific destination while seeing some places of interest along the way) and a circular route returning to the starting point in a specified time (e.g., if the tourist wants to walk around the hotel or a train station).

An algorithm for a fixed pattern noise correction for video sequences with motion objects and/or entire scene based on rows permutation in previous frames and their recurrent accumulation is considered. It is shown that fixed pattern noise is successfully compensated if number of previous frames in sequence more than 300 frames.

The task of the objects identification, counting, and measurement is a huge part of scientific investigations and technological applications. Automated methods using traditional processing such as segmentation, edge detection, and so on represented by available software (e.g. CellProfiler) are not flexible, can be used only with images of high-quality, and in addition require setting a part of parameters by hand. This contribution presents the applying the deep learning method for recognition of HeLa cells expressing green fluorescent protein (EGFP) automatically. We used Cascade Mask R-CNN neural networks which has a ResNeXt backbone and deformable convolutional networks layers. Training dataset contained seven pictures with 5754 labeled cells. Three images with 2469 labeled cells were used as test-dataset. The trained neural network showed mAP=0.4.

the analysis and synthesis of automatic control systems of a thermal object with reference adaptive control models based on the modified MIT rule and the Lyapunov rule using MATLAB-Simulink software is carried out. The developed algorithms have successfully passed laboratory tests on a thermal object and provide the required quality of the transition process in a thermal object.

the paper considers the analysis and synthesis of automatic control systems of a thermal object based on an adaptive fuzzy PID controller and a model of reference adaptive control with the MIT rule using MATLAB-Simulink software. These algorithms have passed laboratory tests at the thermal object and provided the required control parameters.

The article discusses the task of using an artificial neural network to predict the exchange rate. The factors affecting the US dollar to ruble exchange rate are analyzed. Based on the results of the analysis, an option is proposed for using a neural network to predict the exchange rate

The article considers an algorithm for solving the problem of indirectly measuring the distance to a wagon by reducing it to solving the Perspective-2-Point problem. It is shown that on a straight section of a railway track at distances up to 50 meters, an absolute error of measuring the distance to a locomotive or wagon is no more than 1‥2 meters.

The article considers two approaches to solving the problem of multispectral images superimposition in vision systems: based on the search for key points correspondences and preliminary photogrammetric calibration using a universal test object. Shows two variants of the test object design. The advantages and disadvantages of superimposition of the visible and infrared images by two strategies are analyzed.

LPCNet vocoder has recently been presented to TTS community and is now gaining increasing popularity due to its effectiveness and high quality of the speech synthesized with it. In this work, we present a modification of LPCNet that is 1.5x faster, has twice less non-zero parameters and synthesizes speech of the same quality. Such enhancement is possible mostly due to two features that we introduce into the original architecture: the proposed vocoder is designed to generate 16-bit signal instead of 8-bit µ-companded signal, and it predicts two consecutive excitation values at a time independently of each other. To show that these modifications do not lead to quality degradation we train models for five different languages and perform extensive human evaluation.

This article discusses the development of specialized receiving equipment for passive and active methods for the search for hydrocarbon and solid minerals. An option is given for solving the registration of seismic / electric fields using instrumental amplifiers and a 24-bit ADC, since such a solution will minimize distortions introduced in the amplification channels, and match the input impedances, which is most important when registering an electromagnetic field with the use of a grounded electric dipole. It is proposed to build the digital part using a 24-bit sigma-delta ADC and a microcontroller built on the basis of the ARM architecture. 50 Hz industrial interference cancellation is solved using a specialized filter on switched capacitors. Such a solution will allow recording signals with amplitudes of the order of 1 μV.

Myanmar information retrieval (MIR) is the system of searching for the information in a document, searching for documents themselves, and also searching for texts in databases, which is written in Myanmar language. Unfortunately, many problems remain in MIR, such as word segmentation, indexing, semantic ambiguous words and many others. This paper proposes a method of semantic query expansion by using WordNet to resolve semantic ambiguous words problem in MIR. The experimental results showed improvements in evaluating information retrieval potential for the Burmese language with 65.4% precision, 78.9%recall and 70.96% F-measure.

The synthesis of fuzzy quadratic control using Matlab, including the identification of differential equations and the definition of parameters of fuzzy regulator and linear quadratic control algorithms, providing the required quality of the transition process occurring in the laboratory thermal object.