Search
Filters
total: 9369
filtered: 6789
-
Catalog
Chosen catalog filters
displaying 1000 best results Help
Search results for: TIME SERIES CLASSIFICATIONLEARNING SYSTEMSCAPSULE NETWORKSDATA MININGMULTI-HEAD CONVOLUTIONAL NEURAL NETWORKSSIGNAL PROCESSING
-
Statistically efficient smoothing algorithm for time-varying frequency estimation
PublicationThe problem of extraction/elimination of a nonstationary sinusoidal signal from noisy measurements is considered. This problem is usually solved using adaptive notch filtering (ANF) algorithms. It is shown that the accuracy of frequency estimates can be significantly increased if the results obtained from ANF are backward-time filtered by an appropriately designed lowpass filter. The resulting adaptive notch smoothing (ANS) algorithm...
-
Influence of accelerometer signal pre-processing and classification method on human activity recognition
PublicationA study of data pre-processing influence on accelerometer-based human activity recognition algorithms is presented. The frequency band used to filter-out the accelerometer signals and the number of accelerometers involved were considered in terms of their influence on the recognition accuracy. In the test four methods of classification were used: support vector machine, decision trees, neural network, k-nearest neighbor.
-
Optical properties of thin TiO2 film deposited on the fiber optic sensor head
PublicationThe presented study was focused on investigation of the titanium dioxide (TiO2) thin film deposited on the fiber tip. The intention of this investigation was using TiO2 film in the construction of the optical fiber sensor head. In the demonstrated construction TiO2 thin layer was deposited on the tip of a commonly used telecommunication single mode optical fiber (SMF-28) by means of the Atomic Layer Deposition (ALD). Thickness...
-
Dynamically positioned ship steering making use of backstepping method and artificial neural networks
PublicationThe article discusses the issue of designing a dynamic ship positioning system making use of the adaptive vectorial backstepping method and RBF type arti cial neural networks. In the article, the backstepping controller is used to determine control laws and neural network weight adaptation laws. e arti cial neural network is applied at each time instant to approximate nonlinear functions containing parametric uncertainties....
-
Numerical model of human head phantom to ensure dosimetry of dose components for boron neutron capture therapy
PublicationExtremely important aspects of the boron neutron capture therapy are, first of all, administering to the patient a boron compound that selectively reaches the neoplastic cells, and in the second step, the verification of the irradiation process. This paper focuses on the latter aspect, which is the detailed dosimetry of the processes occurring after the reaction of thermal neutrons with the boron-10 isotope. The results of computer...
-
Comparison of the Ability of Neural Network Model and Humans to Detect a Cloned Voice
PublicationThe vulnerability of the speaker identity verification system to attacks using voice cloning was examined. The research project assumed creating a model for verifying the speaker’s identity based on voice biometrics and then testing its resistance to potential attacks using voice cloning. The Deep Speaker Neural Speaker Embedding System was trained, and the Real-Time Voice Cloning system was employed based on the SV2TTS, Tacotron,...
-
Real-time simulation in non real-time environment
PublicationSimulation in real-time is a very useful tool because of didactical and practical benefits. Very important benefit of real-time simulation is a fact that operator’s decision can be taken into account in the same time scale as the real system would work. This enables construction of simulators, and opportunity to test control algorithms in Hardware in The Loop scheme using target industrial equipment. Professional real-time environments...
-
A Conformal Circularly Polarized Series-Fed Microstrip Antenna Array Design
PublicationA conformal circularly polarized series-fed microstrip array design for broadside radiation is presented. The array aperture under design is conformal to a cylindrical surface of a given radius. The approach we present primarily addresses focusing of the circularly polarized major lobe of the conformal array by proper dimensioning of the aperture spacings. The proposed analytical models yield the values of the element spacings...
-
Real-time speech-rate modification experiments
PublicationAn algorithm designed for real-time speech time scale modification (stretching) is proposed, providing a combination of typical synchronous overlap and add based time scale modification algorithm and signal redundancy detection algorithms that allow to remove parts of the speech signal and replace them with the stretched speech signal fragments. Effectiveness of signal processing algorithms are examined experimentally together...
-
A Fast Method of Separation of the Noisy Background from the Head-Cross Section in the Sequence of MRI Scans
PublicationThe paper presents a new method of removing the noisy background from the sequence of magnetic resonance imaging (MRl) scans. The sequence of scans is required in order to monitor a passage of a contrast agent through the brain tissue. The scans contain the noisy head-cross data and also the noisy background data. The latter has to be removed and excluded from a further analysis. It is achieved by applying some basic morphological...
-
Adding Intelligence to Cars Using the Neural Knowledge DNA
PublicationIn this paper we propose a Neural Knowledge DNA based framework that is capable of learning from the car’s daily operation. The Neural Knowledge DNA is a novel knowledge representation and reasoning approach designed to support discovering, storing, reusing, improving, and sharing knowledge among machines and computing devices. We examine our framework for drivers' classification based on their driving behaviour. The experimental...
-
Ordinal Pattern Statistics for RR Intervals during Head-Up Tilt Test in Patients with the History of Vasovagal Syncope
PublicationWe apply ordinal pattern analysis to quantify differences in distribution of patterns of length 3 and 4 in basal state and during head-up tilt test (HUTT) in patients with the history of syncope and positive (HUTT(+)) or negative (HUTT(-)) responses to the test. We identify the patterns related to prevalence of sympathetic or parasympathetic cardiac modulation as well as describe the relations between the response to the test and...
-
Graph Neural Networks and Structural Information on Ionic Liquids: A Cheminformatics Study on Molecular Physicochemical Property Prediction
PublicationIonic liquids (ILs) provide a promising solution in many industrial applications, such as solvents, absorbents, electrolytes, catalysts, lubricants, and many others. However, due to the enormous variety of their structures, uncovering or designing those with optimal attributes requires expensive and exhaustive simulations and experiments. For these reasons, searching for an efficient theoretical tool for finding the relationship...
-
Blood Pressure Estimation Based on Blood Flow, ECG and Respiratory Signals Using Recurrent Neural Networks
PublicationThe estimation of systolic and diastolic blood pressure using artificial neural network is considered in the paper. The blood pressure values are estimated using pulse arrival time, and additionally RR intervals of ECG signal together with respiration signal. A single layer recurrent neural network with hyperbolic tangent activation function was used. The average blood pressure estimation error for the data obtained from 21 subjects...
-
Toward Intelligent Recommendations Using the Neural Knowledge DNA
PublicationIn this paper we propose a novel recommendation approach using past news click data and the Neural Knowledge DNA (NK-DNA). The Neural Knowledge DNA is a novel knowledge representation method designed to support discovering, storing, reusing, improving, and sharing knowledge among machines and computing systems. We examine our approach for news recommendation tasks on the MIND benchmark dataset. By taking advantages of NK-DNA, deep...
-
Comparative study of neural networks used in modeling and control of dynamic systems
PublicationIn this paper, a diagonal recurrent neural network that contains two recurrent weights in the hidden layer is proposed for the designing of a synchronous generator control system. To demonstrate the superiority of the proposed neural network, a comparative study of performances, with two other neural network (1_DRNN) and the proposed second-order diagonal recurrent neural network (2_DRNN). Moreover, to confirm the superiority...
-
Radar and Sonar Imaging and Processing
PublicationThe 21 papers (from 61 submitted) published in the Special Issue “Radar and Sonar Imaging Processing” highlighted a variety of topics related to remote sensing with radar and sonar sensors. The sequence of articles included in the SI dealt with a broad profile of aspects of the use of radar and sonar images in line with the latest scientific trends. The latest developments in science, including artificial intelligence, were used.
-
Times Series Analysis Of Road Safety Trends At The Regional Level In Poland
PublicationThe paper presents possibility of applying a structural times series modeling with explanatory and intervention variables as a tool capable for explaining the changes in the monthly number of fatalities and seriously injured in traffic accidents. The analysis covers regional level and takes into consideration traffic accidents data of two regions: Pomorskie and Warmia-Mazury. In addition short-term forecasts for the two regions...
-
Processing of Satellite Data in the Cloud
PublicationThe dynamic development of digital technologies, especially those dedicated to devices generating large data streams, such as all kinds of measurement equipment (temperature and humidity sensors, cameras, radio-telescopes and satellites – Internet of Things) enables more in-depth analysis of the surrounding reality, including better understanding of various natural phenomenon, starting from atomic level reactions, through macroscopic...
-
ReFlexeNN - the Wearable EMG Interface with Neural Network Based Gesture Classification
PublicationThe electromyographic activity of muscles was measured using a wireless biofeedback device. The aim of the study was to examine the possibility of creating an automatic muscle tension classifier. Several measurement series were conducted and the participant performed simple physical exercises - forcing the muscle to increase its activity accordingly to the selected scale. A small wireless device was attached to the electrodes placed...
-
Modeling of Surface Roughness in Honing Processes by UsingFuzzy Artificial Neural Networks
PublicationHoning processes are abrasive machining processes which are commonly employed to improve the surface of manufactured parts such as hydraulic or combustion engine cylinders. These processes can be employed to obtain a cross-hatched pattern on the internal surfaces of cylinders. In this present study, fuzzy artificial neural networks are employed for modeling surface roughness parameters obtained in finishing honing operations. As...
-
A compact smart sensor based on a neural classifier for objects modeled by Beaunier's model
PublicationA new solution of a smart microcontroller sensor based on a simple direct sensor-microcontroller interface for technical objects modeled by two-terminal networks and by the Beaunier’s model of anticorrosion coating is proposed. The tested object is stimulated by a square pulse and its time voltage response is sampled four times by the internal ADC of microcontroller. A neural classifier based on measurement data classifies the...
-
Web-based 3D processing and dissemination of multibeam sonar data
PublicationThe continuous detailed surveys of the various water bodies over time produce a large and ever-increasing volume and density of underwater sounding data. Three-dimensional data, such as those obtained by multibeam sonar systems, are quite complex to manage, and thus their growing numbers increase the pressure on development of new solutions dedicated to processing them. This paper presents a concept system for web-based dissemination...
-
Implementation of DIFAR Processing in ASW Dipping Sonar
PublicationThis paper presents the implementation of the signal processing algorithm used by buoy DIFAR (Directional Frequency Analysis and Recording), that is fully operational on Polish Navy anti-submarine warfare (ASW) helicopters and ships, applied to dipping sonars for detection and tracking of submarines. The development of the DSP algorithms was a part of the modernization of sonars conducted at the Gdansk University of Technology....
-
Earthquake–induced pounding on response of adjacent structures in series: experimental and numerical study
PublicationPounding between structures in series during earthquakes may cause serious damage in the structur-al elements. The aim of this paper is to show the results of an experimental and numerical study that is focused on pounding between more than two structures which may be called "structures in series". In this study, the shaking table experiments, as well as the numerical analyses, were performed using three tower models including...
-
The effect of processing on the safety and nutritional value
PublicationAgricultural crops as well as animal food raw materials are usually somehow processed before being used by humans. Processing should make them more useful, increase their safety and nutritional value, extend the shelf life, and modify the sensory properties. However, changes and interactions of various compounds in the conditions of processing may generate products toxic or otherwise unsafe for the human organism. The food industry...
-
Explainable AI for Inspecting Adversarial Attacks on Deep Neural Networks
PublicationDeep Neural Networks (DNN) are state of the art algorithms for image classification. Although significant achievements and perspectives, deep neural networks and accompanying learning algorithms have some important challenges to tackle. However, it appears that it is relatively easy to attack and fool with well-designed input samples called adversarial examples. Adversarial perturba-tions are unnoticeable for humans. Such attacks...
-
Optical Sensor Based Gestures Inference Using Recurrent Neural Network in Mobile Conditions
PublicationIn this paper the implementation of recurrent neural network models for hand gesture recognition on edge devices was performed. The models were trained with 27 hand gestures recorded with the use of a linear optical sensor consisting of 8 photodiodes and 4 LEDs. Different models, trained off-line, were tested in terms of different network topologies (different number of neurons and layers) and different effective sampling frequency...
-
UAV Design and Construction for Real Time Photogrammetry and Visual Navigation
PublicationA unmanned aerial vehicles applications in photogrammetry have increased rapidly last years. A fast data gathering and processing in real time in some cases become crucial and desired in some application. In the paper, a real time solution is proposed. A real time photogrammetry from UAV is proposed, where image data are gathered and processed on board UAV and finally reconstructed 3D model and measurements are delivered. The paper...
-
Controlling computer by lip gestures employing neural network
PublicationResults of experiments regarding lip gesture recognition with an artificial neural network are discussed. The neural network module forms the core element of a multimodal human-computer interface called LipMouse. This solution allows a user to work on a computer using lip movements and gestures. A user face is detected in a video stream from a standard web camera using a cascade of boosted classifiers working with Haar-like features....
-
Multibeam sonar data processing for seafloor characterisation
PublicationThe approach to seafloor characterisation was investigated. It relies on calculation of several descriptors (parameters) related to seabed type using three types of multibeam sonar data obtained during seafloor sensing: 1) the grey-level sonar images of seabed, 2) the 3D model of the seabed surface which consist of (x, y, z) points, 3) the set of time domain echo envelopes corresponding to several beams. The proposed method has...
-
Implementation of FDTD-Compatible Green's Function on Graphics Processing Unit
PublicationIn this letter, implementation of the finite-difference time domain (FDTD)-compatible Green's function on a graphics processing unit (GPU) is presented. Recently, closed-form expression for this discrete Green's function (DGF) was derived, which facilitates its applications in the FDTD simulations of radiation and scattering problems. Unfortunately, implementation of the new DGF formula in software requires a multiple precision...
-
Neural Approximators for Variable-Order Fractional Calculus Operators (VO-FC)
PublicationThe paper presents research on the approximation of variable-order fractional operators by recurrent neural networks. The research focuses on two basic variable-order fractional operators, i.e., integrator and differentiator. The study includes variations of the order of each fractional operator. The recurrent neural network architecture based on GRU (Gated Recurrent Unit) cells functioned as a neural approximation for selected...
-
Swapping Space for Time: An Alternative to Time-Domain Interferometry
PublicationYoung's double-slit experiment [1] requires two waves produced simultaneously at two different points in space. In quantum mechanics the waves correspond to a single quantum object, even as complex as a big molecule. An interference is present as long as one cannot tell for sure which slit is chosen by the object. The more we know about the path, the worse the interference. In the paper we show that quantum mechanics allows for...
-
Distributed state estimation using a network of asynchronous processing nodes
PublicationWe consider the problem of distributed state estimation of continuous-time stochastic processes using a~network of processing nodes. Each node performs measurement and estimation using the Kalman filtering technique, communicates its results to other nodes in the network, and utilizes similar results from the other nodes in its own computations. We assume that the connection graph of the network is not complete, i.e. not all nodes...
-
Distributed state estimation using a network of asynchronous processing nodes
PublicationWe consider the problem of distributed state estimation of continuous-time stochastic processes using a~network of processing nodes. Each node performs measurement and estimation using the Kalman filtering technique, communicates its results to other nodes in the network, and utilizes similar results from the other nodes in its own computations. We assume that the connection graph of the network is not complete, i.e. not all nodes...
-
Electromagnetic Modeling of Microstrip Elements Aided with Artificial Neural Network
PublicationThe electromagnetic modeling principle aided withartificial neural network to designing the microwave widebandelements/networks prepared in microstrip technology is proposedin the paper. It is assumed that the complete information is knownfor the prototype design which is prepared on certain substratewith certain thickness and electric permittivity. The longitudinaland transversal dimensions of new design...
-
Modelling and simulation of GPU processing in the MERPSYS environment
PublicationIn this work, we evaluate an analytical GPU performance model based on Little's law, that expresses the kernel execution time in terms of latency bound, throughput bound, and achieved occupancy. We then combine it with the results of several research papers, introduce equations for data transfer time estimation, and finally incorporate it into the MERPSYS framework, which is a general-purpose simulator for parallel and distributed...
-
Discrete-time estimation of nonlinear continuous-time stochastic systems
PublicationIn this paper we consider the problem of state estimation of a dynamic system whose evolution is described by a nonlinear continuous-time stochastic model. We also assume that the system is observed by a sensor in discrete-time moments. To perform state estimation using uncertain discrete-time data, the system model needs to be discretized. We compare two methods of discretization. The first method uses the classical forward Euler...
-
Discrete-time estimation of nonlinear continuous-time stochastic systems
PublicationIn this paper we consider the problem of state estimation of a dynamic system whose evolution is described by a nonlinear continuous-time stochastic model. We also assume that the system is observed by a sensor in discrete-time moments. To perform state estimation using uncertain discrete-time data, the system model needs to be discretized. We compare two methods of discretization. The first method uses the classical forward Euler...
-
Real-time speech streching for supporting hearing impaired schoolchildren
PublicationA study of time scale modification algorithms applied to support hearing impaired schoolchildren is presented. Variety of algorithms are considered, namely: overlap-and add, two variations of synchronous overlapand- add, and the phase vocoder. Their effectiveness as well as real-time processing capabilities are examined.
-
Neural network based algorithm for hand gesture detection in a low-cost microprocessor applications
PublicationIn this paper the simple architecture of neural network for hand gesture classification was presented. The network classifies the previously calculated parameters of EMG signals. The main goal of this project was to develop simple solution that is not computationally complex and can be implemented on microprocessors in low-cost 3D printed prosthetic arms. As the part of conducted research the data set EMG signals corresponding...
-
A fast time-frequency multi-window analysis using a tuning directional kernel
PublicationIn this paper, a novel approach for time-frequency analysis and detection, based on the chirplet transform and dedicated to non-stationary as well as multi-component signals, is presented. Its main purpose is the estimation of spectral energy, instantaneous frequency (IF), spectral delay (SD), and chirp rate (CR) with a high time-frequency resolution (separation ability) achieved by adaptive fitting of the transform kernel. We...
-
Diagnosis of damages in family buildings using neural networks
PublicationThe article concerns a problem of damages in family buildings, which result from traffic-induced vibrations. These vibrations arise from various causes and their size is influenced by many factors. The most important is the type of a road, type and weight of vehicles that run on the road, type and condition of the road surface, the distance from the house to the source of vibrations and many others which should be taken into account....
-
Survival time prognosis under a Markov model of cancer development
PublicationIn this study we look at a breast cancer data set of women from Pomerania region collected in year 1987-1992 in the Medical University of Gdańsk. We analyze the clinical risk factors in conjunction with Markov model of cancer development. We evaluate Artificial Neural Network (ANN) survival time prediction via a simulation study.
-
A new optimal algorithm for a time-dependent scheduling problem
PublicationIn this article a single machine time-dependent scheduling problem with total completion time criterion is considered. There are n given jobs j_1, ..., j_n and the processing time pi of the i-th job is given by p_i = 1 + b_is_i, where si is the starting time of the i-th job, i = 1, ..., n. If all jobs have different and non-zero deterioration rates and bi > bj => bi >= (b_min+1)/(b_min) b_j + 1/b_min, where b_min = min{b_i}, then...
-
Design and Analysis of Artificial Neural Network (ANN) Models for Achieving Self-Sustainability in Sanitation
PublicationThe present study investigates the potential of using fecal ash as an adsorbent and demonstrates a self-sustaining, optimized approach for urea recovery from wastewater streams. Fecal ash was prepared by heating synthetic feces to 500 °C and then processing it as an adsorbent for urea adsorption from synthetic urine. Since this adsorption approach based on fecal ash is a promising alternative for wastewater treatment, it increases...
-
An Analog Sub-Miliwatt CMOS Image Sensor With Pixel-Level Convolution Processing
PublicationA new approach to an analog ultra-low power medium-resolution vision chip design is presented. The prototype chip performs low-level image processing algorithms in real time. Only a photo-diode, MOS switches and two capacitors are used to create an analog processing element (APE) that is able to realize any convolution algorithm based on a full 3x3 kernel. The proof-of-concept circuit is implemented in 0.35 µm CMOS technology,...
-
Effect of head-to-tail cyclization on conformation of histatin-5
Publication -
The SPM Scanner Head Based on Piezoelectric Unimorph Disc
Publication