Abstract The paper considers the generation of effective quantization scales which meet any quality criteria under any constrains. The analysis of work of available methods allows a concept of a heuristic iterative algorithm for generating near-optimal quantization scales. The concept uses a uniform scale as an initial approach followed by an iterative target-criterion-optimizing recalculation of quantization

Abstract Lidar observations of aerosol and ozone, carried out at Siberian Lidar Station (SLS) of Institute of Atmospheric Optics, Siberian Branch, Russian Academy of Sciences in Tomsk (56°29′ N; 85°3′ E), showed the presence of stratospheric aerosol layers over Tomsk during winter 2017–2018, signs of descending air masses, and deficit of ozone. The Aura OMI/MLS observations indicated that in December-January

Abstract The previously developed n-vicinity method [3] allows us to calculate accurately critical values of inverse temperatures for Ising models with short-range interaction. We generalize the method to the case of long-range interactions in spin systems and obtain theoretical formulas for the inverse temperatures in terms of the spin interaction constants. The comparison of our theoretical estimates

Abstract The paper proposes a new approach to constructing a decentralized hierarchical network of modular type for solving complex problems in the paradigm of training artificial neural networks (ANN) with a teacher. The main idea of the proposed approach to solving the problem is the formation of a system model in the form of a multilevel hierarchical network, consisting of typical autonomous modules

Abstract Rumor detection and stance classification are specialized areas in the field of Information Retrieval and Natural Language Processing. Judging the stances of public response is viewed as an important preceding step of rumor veracity prediction. In this paper we develop a reinforcement learning-based multi-task learning for rumor early detection (RL_MT_RED), which formulates the closely correlated

Abstract Student performance prediction is a primary goal for educational institutions to improve future performance of students. Early prediction of student’s future performance is critical and it may result interventions to students. Researchers focused in this field and attempt to give a better prediction at early stage. Previously existing algorithms namely fuzzy logic was implemented. However

Abstract An algorithm of software complex is developed to process data of lidar gas analysis in IR wavelength range for a horizontal path of atmospheric sensing. The complex comprises: the module of writing data from atmospheric lidar sensing, module of absorption cross-section calculation, and module of methane concentration profile retrieval. The software complex for processing lidar gas analysis

Abstract The article presents the results of research on development of a method for calculating human operator recognition probability of objects noisy images generated by optoelectronic surveillance devices. The proposed method is based on a visual system mathematical model which takes into account features of pre-processing images performed in human eyes as well as at stages of primary and secondary

Abstract We discuss a random search algorithm with self-learning designed for solving a problem of training of feedforward neural networks and compare it with gradient algorithms for neural network training with regard to criteria of accuracy and computational complexity.

Abstract This article describes the development of a method for improving the security of fingerprint identification systems. It is based on the calculation of the polarimetric parameters of biological tissue using Mueller matrices. This method was tested on samples made from gelatin, which have a copy of the papillary pattern and mimic real tissue. It is shown that the proposed method is promising

Abstract Image segmentation can extract valuable information from images and has very important practical significance. In this paper, the application of Convolutional Neural Network (CNN) in image processing is studied. Full Convolutional Network (FCN) is used to improve the accuracy of image feature extraction and Visual Geometry Group-16 (VGG-16) is improved. In order to further improve the accuracy

Abstract Accurate and timely traffic forecasting plays an important role in the development of intelligent transportation systems (ITS). Traffic data are the main information source for various tasks solved as part of the ITS, including traffic management, urban planning, route guidance, and others. Due to the spatial and temporal nonlinearity and complexity of traffic flow, traffic forecasting problem

Abstract As a high performance method for various image processing tasks, deep convolutional neural networks (CNNs) have reached impressive performances and absorbed considerable attention in the last few years. However, object classification on small size datasets for which a limited number of training images is available is still considered as an open problem. In this paper, we investigate a new

Abstract This paper is concerned with the generalized finite-time stability, boundedness and stabilization for fractional-order memristive neural networks (FMNNs) with the fractional-order 0 < α < 1. Under the fractional-order Filippov differential inclusion frame, FMNNs are modelled as a fractional-order differential equation with discontinuous right-hand. Based on the topological degree property

Abstract One of the ways to join the connectionist approach and the symbolic paradigm is Tensor Product Variable Binding. It was initially devoted to building distributed representation of recursive structures for neural networks to use it as the input. Structures are an essential part of both formal and natural languages and appear in syntactic trees, grammar, semantic interpretation. A human mind

Abstract Health monitoring of an aero-engine assumes importance in the light of primary requirements of flight safety and reliability. This paper proposes a novel, simple method for monitoring aircraft engine health using Whale Optimization Algorithm based Artificial Neural Network (WOANN) technique, for analyzing the data downloaded from the health and usage monitoring system (HUMS) of a military

Abstract The neural network-based approach to the classification of implicit emotional components in ordinary speech is considered. Mel-frequency cepstral coefficients were used as feature vectors, and the multilayer perceptron with one hidden layer was used as the classifier. It was shown that the neural-network system developed is able to classify these kinds of speech with the accuracy up to 99%

Abstract The results of comparative spectral assessment of circumpulpal and mantle dentin and possibility of their further use for making bone substitute materials in dental surgery and implantology are presented in the work. The research was made using the Raman spectroscopy method. It is shown that the researched biological subjects have similar spectral composition and tooth dentin can be eventually

Abstract We analyze a finite one-dimensional Ising system with periodic boundary conditions taking into account an arbitrary long-range interaction. We examine a discrete spectrum of eigenvalues of the spin connection matrix and a spectrum density of a continuous distribution obtained in the limit \(L \to \infty \) (L is the linear size of the system). We apply our results to particular cases of long-range

Abstract We propose a new information technology for mapping macular photocoagulation patterns in the laser treatment of diabetic retinopathy and evaluating the treatment strategy efficacy. The technology relies on processing optical coherence tomography (OCT) data, fundus image segmentation, mapping photocoagulation plans, and development of prognostic indicators of the laser photocoagulation treatment

Abstract In this paper, we analyze effective methods of multi-label classification of image sets in development of visual recommender systems. We propose a two-step algorithm, which at the first step performs fine-tuning of a convolutional neural network for extraction of visual features. At the second stage, the algorithm concatenates the obtained feature vectors of each image from the input set into

Abstract Laser scanning technique is an important area of the optical and laser technology, which makes the access of 3D individual tree information becomes available. In order to deal with the biomass and structure estimation of the urban forest, many algorithms have been developed for 3D point clouds to extract individual tree information, including tree counts, tree locations, branching structure

Abstract This review offers the reader some of the achievements of modern optical metrology. Over the past decades, it has become possible to make a leap in the basic approaches of metrology from the nano to the femto, approaching the pico level of measurements. Control of nano (micro) particle motion by an optical field and their use for testing complex optical fields; ultra-precise determination

Abstract Data sharing in the Peer to Peer (P2P) networks became an important function in the trustworthy computing. Secure and load balancing control in file sharing is vital to enhance the overall performance of P2P file sharing system. In literature many methods of load balancing control and security control have been used but it is not able to attain the best results in P2P networks. Hence, in this

Abstract Observation of films of biological liquids allows finding markers of diseases such as diabetes or pneumonia. This may allow switching to more cost-effective diagnostic methods based on the use of relatively cheap commodity hardware required for optical microscopy. For this propose cuneiform dehydration of biological fluids is a promising method of medical diagnosis based on the study of structures

Abstract Among the tasks solved by artificial neural networks are the tasks of analyzing objects on the images of the underlying Earth’s surface, obtained by the on-board equipment of unmanned aerial vehicle (UAV). For the solution of such problems, the convolutional neural networks (CNN), operating semantic segmentation of the received image, are widely used. In this case, the designer of such networks

Abstract— The performance of RAMAN-TDSGF hybrid optical amplifier (HOA) has been examined in the context of gain and noise figure (NF) for 400 × 10 Gbps super dense wavelength division multiplexing (SD-WDM) system. Characteristics of this novel HOA has also compared with RAMAN-EDFA-RAMAN, RAMAN-EDFA and RAMAN-RAMAN HOA for the range of 1530 nm to 1665, and 1530 nm to 1586 nm respectively. Best rating

Abstract— In this article, we develop method of linear and exponential quantization of neural network weights. We improve it by means of maximizing correlations between the initial and quantized weights taking into account the weight density distribution in each layer. We perform the quantization after the neural network training without a subsequent post-training and compare our algorithm with linear

Abstract A new method for determining the chirp of a femtosecond broadband IR laser pulse at a central wavelength of 2.5 μm, based on the generation of a spectral supercontinuum in the field of a spectrally limited femtosecond laser pulse, propagating in a single mode fiber based on chalcogenide glass, and subsequent noncollinear generation of radiation at summary frequency by spectrally limited and

Abstract— Last few decades there are lots of changes in Southeast Asia land cover due to development, industrialization, increasing population, socio-economic activities, and food demands. This research work analysis the maximum likelihood supervised classification approach for Southeast Asia land cover mapping and changes at 30m resolution from 2010 to 2019 using Landsat 8 OLI (Operational Land Imager)

Abstract The research considers the minimization of spin glass energy via learning and evolution. The Sherrington-Kirkpatrick spin-glass model is used. A population of autonomous agents is considered. The genotype and phenotype of each agent are chains consisting of a great number of spins. The energy of spin glasses is minimized through learning and evolution of agents. The genotypes of agents are

Abstract— this paper describes an algorithm for classification of roof materials using aerial photographs. Main advantages of the algorithm are proposed methods to improve prediction accuracy. Proposed methods includes: method of converting ImageNet weights of neural networks for using multi-channel images; special set of features of second level models that are used in addition to specific predictions

Abstract The article presents the results of research on the development of a method for calculating detection probability of noisy objects images by a human. The proposed calculation method are based on the visual system models, which take into account the features of images pre-processing carried out in the human eyes, as well as at the stages of primary and secondary processing performed in the

Abstract In this article the image processing of workflow from raw camera frames is detailed. The results of the visualization and quantitative analysis of the images from the rhodamine 6 G solution obtained using total internal reflection fluorescence microscopy are presented. The main stages for image processing of fluorescent molecules are considered. The following actions are described: acquisition

Abstract— From the last decade, the use of internet and its growth is continuously increasing. Similarly, numbers of services are coming out along with the internet and it is being used for providing facilities to human beings. Wireless sensor have been used for various application such as fire safety, military application, petroleum industry, security system, monitoring and environmental condition

Abstract— The paper investigates the use of generative adversarial networks (GAN) for intentional modification of Earth remote sensing data. A generative neural network that includes a special subnet for object boundary inpainting is considered. The network comprises two GAN: the first completes the object boundary, and the second repaints blank areas. Actual remote sensing data are used to test the

Abstract Chromatic Dispersion Compensation (CDC) using Fiber Bragg Grating (FBG) techniques in Dense Wavelength Division Multiplexing (DWDM) system is presented in this article. Pulse broadening effects on the transmitted signal can be limited by Dispersion Compensation (DC) mechanisms. To overcome this problem and improve the system performance, a suggested system for 10 Gbps using Chirped Fiber Bragg

AbstractAn experimental study of laser ablation on the samples that comprise nanofilms of hafnium and zirconium dioxides on the surface of glass and silicon, and obtaining parameters of pulsed laser ablation in this work, provide new data describing the ablation mechanism. The samples were exposed to pulsed radiation from an Nd3+:YAG laser in order to measure the threshold energy density of laser ablation

AbstractMathematical simulations of invariant optical-digital correlator operation are performed with the LC SLM used for the input images display and for display of the correlation filter holograms. Different phase dependences on amplitude are considered, particularly the measured dependence of the LC SLM HoloEye LC 2002. A method for optimization of correlation filters to eliminate the object recognition

AbstractStructure of the fiber-optic recirculating loop memory with periodic 2R-regeneration of information flow in the optical range is designed. The modified optical fiber loop mirror for 2R-regeneration of wavelength division multiplexed return-to-zero signals is proposed. On the basis of the proposed mathematical model, using numerical modeling, the efficiency of the restoration of optical information

AbstractA method for visualization of volumetric defects in a ZnGeP2 single-crystal by digital holography method using strontium vapor laser radiation is proposed. The possibility of obtaining a volume distribution of defects with dimensions of ≥15–20 μm and their identification in a crystal is shown. The identification in a ZnGeP2 single-crystal of such volumetric defects as growth bands and needle

AbstractThe paper deals with the algorithms of optimal quantization and adaptive interpolation in interpolative and hierarchical compression of multidimensional signals. An uneven quantization scale optimization algorithm is invented to tackle the unknown number of quantization levels given soft requirements for the optimization criterion. An important instance of applying the quantizer to the compression

AbstractThis paper is concerned with the global Mittag-Leffler stability (GMLS) and global finite-time stability (GFTS) for fractional Hopfield neural networks (FHNNs) with Hölder neuron activation functions subject to nonlinear growth. Firstly, four functions possessing convexity are proposed, which can guarantee that four formulas with respect to the fractional derivative are established. Correspondingly

AbstractIn this paper, a parallel SSD (Single Shot MultiBox Detector) fusion network based on inverted residual structure (IR-PSN) is proposed to solve the problems of the lack of extracted feature information and the unsatisfactory effect of small object detection by deep learning. Firstly, the Inverted Residual Structure (IR) is adopted into the SSD network to replace the pooling layer. The improved

AbstractTraining Convolutional Neural Networks that do well in one-shot learning settings can have wide range of impacts on real-world datasets. In this paper, we explore an adversarial training method that learns a Siamese neural network in an end-to-end fashion for two models—ConvNets model that learns image embeddings from input image pair, and head model that further learns the distance between

AbstractIn this paper we discuss image processing algorithms for analysis of structures in the films of biological liquids obtained by the cuneiform dehydration method. In medical diagnostics the cuneiform dehydration may be considered as a promising method of biological liquids studies. Main types of structures in the biological liquid films are presented. Today the methods of interpretation and algorithms

AbstractThe problem of neural network synthesis on the precedents is addressed. The aim of the study is to create methods for radial-basis neural network model constructing having high levels of generalization and accuracy, which do not require user participation in the process of model building. The method of decision tree transforming into a neural network model is proposed. For a given sample, a

AbstractThe goal of the study is to increase the computation efficiency of the face recognition that uses feature vectors to describe facial images on photos and videos. These high-dimensional feature vectors are nowadays produced by convolutional neural networks. The methods to aggregate the features generated for each video frame are used to process the video sequences. A novel hierarchical recognition

AbstractAlthough Deep Neural Networks (DNNs) have performed great success in machine learning domain, they usually show poorly on few-shot learning tasks, where a classifier has to quickly generalize after getting very few samples from each class. A Model-Agnostic Meta Learning (MAML) model, which is able to solve new learning tasks, only using a small number of training data. A MAML model with a Convolutional

AbstractWith the rapid development of Internet technology, the data information on the network has also increased in an unprecedented amount. However, the emergence of this phenomenon makes it difficult for network users to find information of great value to themselves in the massive data. Therefore, score prediction is very important. It is an urgent task to establish an effective and practical scoring

AbstractIn social media, organizing friendship relationships is difficult since the more number of increasing users in Online Social Networks (OSN). To overcome this challenge, users in OSN heavily depend on grouping which is considered to be advantageous but at the same time found to be more cumbersome. More recently, recommender system and novel data clustering algorithm have been presented in OSN

AbstractIn the present paper, we present a multi-start method with dynamic cutting of “unpromising” starts of locally optimal algorithms for solving continuous unconditional optimization problems. We also describe our results of testing of the proposed method for solving problems of feed-forward neural network training.

AbstractAmong a number of problems in the behavior planning of an unmanned vehicle the central one is movement in difficult areas. In particular, such areas are intersections at which direct interaction with other road agents takes place. In our work, we offer a new approach to train of the intelligent agent that simulates the behavior of an unmanned vehicle, based on the integration of reinforcement

AbstractRadial Basis Function Neural Networks (RBFNNs) have been widely used for classification and function approximation tasks. So, it is worthy to try improving and developing new learning algorithms for RBFNNs in order to get better results. This paper presents a new learning method for RBFNNs. Hence, an improved learning algorithm for center adjustment of RBFNNs using Harmony search (HS) algorithm

AbstractPhysical structures changing their resistance in operation can serve as a basis for making elements of neural networks (synapses, neurons, etc.). The processes inducing changes of resistance are rather complicated and cannot be described readily. To demonstrate the potential of this sort of variable resistors it is possible to substitute a complex physical system by a simple mathematical model

AbstractData encryption is becoming increasingly relevant with the development of digital technologies. A particularly promising direction is the development of encryption methods based on optical transformations. Fractional Fourier Transform is a well-known method of encoding data, especially graphic. However, an assessment of the method’s resistance to unauthorized data decryption has not been carried

AbstractIn this paper, the global Mittag-Leffler stability of fractional Hopfield neural networks (FHNNs) with \(\delta \)-inverse hölder neuron activation functions are considered. By applying the Brouwer topological degree theory and inequality analysis techniques, the proof of the existence and uniqueness of equilibrium point is addressed. By constructing the Lure’s Postnikov-type Lyapunov functions

AbstractIn Free Space Optical Communication (FSO), the optical signal from the laser source severely affects while travelling through free space atmospheric channel due to scattering, absorption and other effects of atmospheric turbulence conditions. This degrades the performance of FSO communication. In this article, we have generated the hollow beam from the laser output by the inverse axicon lens

AbstractTo reduce random access memory (RAM) requirements and to increase speed of recognition algorithms we consider a weight discretization problem for trained neural networks. We show that an exponential discretization is preferable to a linear discretization since it allows one to achieve the same accuracy when the number of bits is 1 or 2 less. The quality of the neural network VGG-16 is already

AbstractThe problem of image understanding currently attracts considerable attention of researchers, since its solution is critically important for a significant number of applied problems. Among the most critical components of this problem is the semantic segmentation of images, which provides a classification of objects on the image at the pixel level. One of the applied problems in which semantic