Filters
total: 10530
-
Catalog
- Publications 8653 available results
- Journals 154 available results
- Conferences 62 available results
- Publishing Houses 2 available results
- People 140 available results
- Inventions 1 available results
- Projects 5 available results
- e-Learning Courses 43 available results
- Events 7 available results
- Open Research Data 1463 available results
displaying 1000 best results Help
Search results for: LOMBARD EFFECT, SPEECH DETECTION, NOISE SIGNAL, SELF-SIMILARITY MATRIX, CONVOLUTIONAL NEURAL NETWORK
-
1D convolutional context-aware architectures for acoustic sensing and recognition of passing vehicle type
PublicationA network architecture that may be employed to sensing and recognition of a type of vehicle on the basis of audio recordings made in the proximity of a road is proposed in the paper. The analyzed road traffic consists of both passenger cars and heavier vehicles. Excerpts from recordings that do not contain vehicles passing sounds are also taken into account and marked as ones containing silence....
-
Application of Complementary Signals in Built-In Self Testers for Mixed-Signal Embedded Electronic Systems
PublicationThis paper concerns the implementation of shape-designed complementary signals (CSs), matched to the frequency characteristic of the circuit under test, in built-in self testers (BISTs), dedicated to mixed-signal embedded electronic systems for testing their analog sections. The essence of the proposed method and solution of CS BIST is low-cost realization on the base of hardware and software resources of microcontrollers used...
-
Neural Network-Based Sequential Global Sensitivity Analysis Algorithm
PublicationPerforming global sensitivity analysis (GSA) can be challenging due to the combined effect of the high computational cost, but it is also essential for engineering decision making. To reduce this cost, surrogate modeling such as neural networks (NNs) are used to replace the expensive simulation model in the GSA process, which introduces the additional challenge of finding the minimum number of training data samples required to...
-
Noise as stimulus signal for digital circuits noise immunity measurementand simulation
PublicationWyniki pomiarów podatności na zakłócenia układów cyfrowych istotnie zależą od parametrów sygnału stymulujacego. Przedstawiono metodę pomiaru i symulacji dynamicznej odporności na zakłócenia z zastosowaniem sygnału szumu białego o różnej szerokości pasma częstotliwości. Modele symulacyjne dla sygnału pobudzającego (gaussowskiego szumu białego) zarówno szerokopasmowego jak i wąskopasmowego opracowano w środowisku programistycznym...
-
Elimination of clicks from archive speech signals using sparse autoregressive modeling
PublicationThis paper presents a new approach to elimination of impulsivedisturbances from archive speech signals. The proposedsparse autoregressive (SAR) signal representation is given ina factorized form - the model is a cascade of the so-called formantfilter and pitch filter. Such a technique has been widelyused in code-excited linear prediction (CELP) systems, as itguarantees model stability. After detection of noise pulses usinglinear...
-
Comparison of single best artificial neural network and neural network ensemble in modeling of palladium microextraction
Publication -
Effect of illumination on noise of silicon solar cells
PublicationTransport and noise properties of silicon solar cells in darkness and under illumination have been studied. The measurements were carried out for both reverse and forward bias of the device. The changes in the sample behaviour are due to the variation of PN junction dynamic resistance and, also, the occurence of both 1/f and generation-recombination noise components
-
Improvement of glass break acoustic signal detection via application of wavelet packet decomposition
PublicationThe main subject of the authors' research are non-contact methods of glass break detection based on analysis of the acoustic signal generated during the event. This problem has essential meaning for modern cost- effective alarm systems, particularly those installed into big buildings. The main difficulties of the matter are: transient character of the signal, great number of similar sounds (false signals, mainly accidental glass...
-
A Comprehensive Analysis of Deep Neural-Based Cerebral Microbleeds Detection System
PublicationMachine learning-based systems are gaining interest in the field of medicine, mostly in medical imaging and diagnosis. In this paper, we address the problem of automatic cerebral microbleeds (CMB) detection in magnetic resonance images. It is challenging due to difficulty in distinguishing a true CMB from its mimics, however, if successfully solved it would streamline the radiologists work. To deal with this complex three-dimensional...
-
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...
-
Speech Intelligibility Measurements in Auditorium
PublicationSpeech intelligibility was measured in Auditorium Novum on Technical University of Gdansk (seating capacity 408, volume 3300 m3). Articulation tests were conducted; STI and Early Decay Time EDT coefficients were measured. Negative noise contribution to speech intelligibility was taken into account. Subjective measurements and objective tests reveal high speech intelligibility at most seats in auditorium. Correlation was found between...
-
Development of a tropical disease diagnosis system using artificial neural network and GIS
PublicationExpert systems for diagnosis of tropical diseases have been developed and implemented for over a decade with varying degrees of success. While the recent introduction of artificial neural networks has helped to improve the diagnosis accuracy of such systems, this aspect is still negatively affected by the number of supported diseases. A large number of supported diseases usually corresponds to a high number of overlapping symptoms,...
-
On the EM algorithm for the estimation of speech AR parameters in noise
Publication -
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...
-
New Two-center Ellipsoidal Basis Function Neural Network for Fault Diagnosis of Analog Electronic Circuits
PublicationIn the paper a new fault diagnosis-oriented neural network and a diagnostic method for localization of parametric faults in Analog Electronic Circuits (AECs) with tolerances is presented. The method belongs to the class of dictionary Simulation Before Test (SBT) methods. It utilizes dictionary fault signatures as a family of identification curves dispersed around nominal positions by component tolerances of the Circuit Under Test...
-
Characterizing the Scalability of Graph Convolutional Networks on Intel® PIUMA
PublicationLarge-scale Graph Convolutional Network (GCN) inference on traditional CPU/GPU systems is challenging due to a large memory footprint, sparse computational patterns, and irregular memory accesses with poor locality. Intel’s Programmable Integrated Unffied Memory Architecture (PIUMA) is designed to address these challenges for graph analytics. In this paper, a detailed characterization of GCNs is presented using the Open-Graph Benchmark...
-
Benchmarking Deep Neural Network Training Using Multi- and Many-Core Processors
PublicationIn the paper we provide thorough benchmarking of deep neural network (DNN) training on modern multi- and many-core Intel processors in order to assess performance differences for various deep learning as well as parallel computing parameters. We present performance of DNN training for Alexnet, Googlenet, Googlenet_v2 as well as Resnet_50 for various engines used by the deep learning framework, for various batch sizes. Furthermore,...
-
The noise-induced harmful effect assessment based on the properties of the human hearing system
PublicationA new way of assessment of noise-induced harmful effects on human hearing system is presented in the paper. The method takes into consideration properties of the human hearing system. The pro-posed method determines the cumulative impact on hearing system produced by the excessive noise. Based on the predicted effects of the noise exposure, the new types of noise indicators were developed. The evaluation of these indicators was...
-
Electrical and noise responses of graphene back-gated field-effect transistors enhanced by UV light for organic vapors sensing
Open Research DataBack-gated field-effect transistors with graphene channels (GFETs) were investigated toward organic vapors sensing. Two methods were used for sensing experiments including DC characteristics measurements and fluctuation-enhanced sensing by low-frequency noise studies. The data set consists of raw and modified data on GFET responses to acetonitrile,...
-
Neural networks and deep learning
PublicationIn this chapter we will provide the general and fundamental background related to Neural Networks and Deep Learning techniques. Specifically, we divide the fundamentals of deep learning in three parts, the first one introduces Deep Feed Forward Networks and the main training algorithms in the context of optimization. The second part covers Convolutional Neural Networks (CNN) and discusses their main advantages and shortcomings...
-
A method of self-testing of analog circuits based on fully differential op-amps with theTCBF classifier
PublicationA new approach of self-testing of analog circuits based on fully differential op-amps of mixed-signal systems controlled by microcontrollers is presented. It consists of a measurement procedure and a fault diagnosis procedure. We measure voltage samples of a time response of a tested circuit on a stimulation of a unit step function given at the common-mode reference voltage input of the op-amp. The fault detection and fault localization...
-
Self-healing and shape memory polymers
e-Learning CoursesThe course "Self-Healing and Shape Memory Polymers" is a specialized elective designed for Chemical Engineering PhD students. This advanced course delves into the innovative realms of self-healing materials and shape memory polymers (SMPs), representing a significant breakthrough in material science and engineering. Throughout the course, students learn about the problems of failure of polymeric materials and explore the fundamental...
-
Simulation of parallel similarity measure computations for large data sets
PublicationThe paper presents our approach to implementation of similarity measure for big data analysis in a parallel environment. We describe the algorithm for parallelisation of the computations. We provide results from a real MPI application for computations of similarity measures as well as results achieved with our simulation software. The simulation environment allows us to model parallel systems of various sizes with various components...
-
Monitoring Regenerative Heat Exchanger in Steam Power Plant by Making Use of the Recurrent Neural Network
PublicationArtificial Intelligence algorithms are being increasingly used in industrial applications. Their important function is to support operation of diagnostic systems. This paper pesents a new approach to the monitoring of a regenerative heat exchanger in a steam power plant, which is based on a specific use of the Recurrent Neural Network (RNN). The proposed approach was tested using real data. This approach can be easily adapted to...
-
Piotr Szczuko dr hab. inż.
PeoplePiotr Szczuko received his M.Sc. degree in 2002. His thesis was dedicated to examination of correlation phenomena between perception of sound and vision for surround sound and digital image. He finished Ph.D. studies in 2007 and one year later completed a dissertation "Application of Fuzzy Rules in Computer Character Animation" that received award of Prime Minister of Poland. His interests include: processing of audio and video, computer...
-
Transient detection algorithms for speech coding applications
Publication -
Influence of modulation detection threshold on speech intelligibility
Publication -
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,...
-
Intra-subject class-incremental deep learning approach for EEG-based imagined speech recognition
PublicationBrain–computer interfaces (BCIs) aim to decode brain signals and transform them into commands for device operation. The present study aimed to decode the brain activity during imagined speech. The BCI must identify imagined words within a given vocabulary and thus perform the requested action. A possible scenario when using this approach is the gradual addition of new words to the vocabulary using incremental learning methods....
-
Design of Microstrip UWB Balun Using Quasi-TEM Approach Aided by the Artificial Neural Network
PublicationThe design procedure for UWB balun realized in the microstrip technology is proposed in the paper. The procedure applies Artificial Neural Network which corrects the dimensions of the approximate design found by appropriate scaling of the dimensions of the prototype. The scale coefficients for longitudinal and transverse dimensions of microstrip lines are determined from electromagnetic modeling based on transmission line equations....
-
Iterative Global Sensitivity Analysis Algorithm with Neural Network Surrogate Modeling
PublicationGlobal sensitivity analysis (GSA) is a method to quantify the effect of the input parameters on outputs of physics-based systems. Performing GSA can be challenging due to the combined effect of the high computational cost of each individual physics-based model, a large number of input parameters, and the need to perform repetitive model evaluations. To reduce this cost, neural networks (NNs) are used to replace the expensive physics-based...
-
Text-mining Similarity Approximation Operators for Opinion Mining in BI tools
PublicationThe concept of the Text-mining Similarity Approximation Operators for Opinion Mining as extensions to Natural Language Interface Database is defined. The new operators: “keywords of” dimension; subsetting operator “about C is q”; aggregation operator “by similar C” are proposed. These operators are based on the Latent Semantic Analysis and Social Network Analysis
-
A New Direct-Sequence Spread Spectrum Signal Detection Method for Underwater Acoustic Communications in Shallow-Water Channel
PublicationDirect-Sequence Spread Spectrum (DSSS) is one of the modulation and coding techniques used in Underwater Acoustic Communication (UAC) systems for reliable data transmision even at low signal levels. However, in a shallow water channel, there is a strong multipath propagation which causes a phase fluctuation of the received signal, affecting the performance of the spread-spectrum system. The article presents a differential method...
-
Modeling the effect of electric vehicles on noise levels in the vicinity of rural road sections
PublicationNumerous European countries experience a steady increase in the share of electric (EV) and hybrid electric (HEV) vehicles in the traffic stream. These vehicles, often referred to as low- or zero-emission vehicles, significantly reduce air pollution in the road environment. They also have a positive effect on noise levels in city centers and in the surroundings of low-speed roads. Nevertheless, issues related to modeling noise from...
-
Artificial neural network based fatigue life assessment of friction stir welding AA2024-T351 aluminum alloy and multi-objective optimization of welding parameters
PublicationIn this paper, the fracture behavior and fatigue crack growth rate of the 2024-T351 aluminum alloy has been investigated. At first, the 2024-T351 aluminum alloys have been welded using friction stir welding procedure and the fracture toughness and fatigue crack growth rate of the CT specimens have been studied experimentally based on ASTM standards. After that, in order to predict fatigue crack growth rate and fracture toughness,...
-
Marek Blok dr hab. inż.
PeopleMarek Blok in 1994 graduated from the Faculty of Electronics at Gdansk University of Technology receiving his MSc in telecommunications. In 2003 received Ph.D. and in 2017 D.Sc. in telecommunications from the Faculty of Electronics, Telecommunications and Informatics of Gdańsk University of Technology. His research interests are focused on application of digital signal processing in telecommunications. He provides lectures, laboratory...
-
A method of RTS noise identification in noise signals of semiconductor devices in the time domain
PublicationIn the paper a new method of Random Telegraph Signal (RTS) noise identification is presented. The method is based on a standardized histogram of instantaneous noise values and processing by Gram-Charlier series. To find a device generating RTS noise by the presented method one should count the number of significant coefficients of the Gram-Charlier series. This would allow to recognize the type of noise. There is always one (first)...
-
Deep neural networks for data analysis
e-Learning CoursesThe aim of the course is to familiarize students with the methods of deep learning for advanced data analysis. Typical areas of application of these types of methods include: image classification, speech recognition and natural language understanding. Celem przedmiotu jest zapoznanie studentów z metodami głębokiego uczenia maszynowego na potrzeby zaawansowanej analizy danych. Do typowych obszarów zastosowań tego typu metod należą:...
-
A random signal generation method for microcontrollers with DACs
PublicationA new method of noise generation based on software implementation of a 7-bit LFSR based on a common polynomial PRBS7 using microcontrollers equipped with internal ADCs and DACs and a microcontroller noise generator structure are proposed in the paper. Two software applications implementing the method: written in ANSI C and based on the LUT technique and written in AVR Assembler are also proposed. In the method the ADC results are...
-
Substrate noise-aware floorplanning for mixed-signal SOCs.
PublicationOpisano nową metodę projektowania systemów na chipie z uwzglednieniem minimalizacji szumu podłożowego. Optymalizacje rozmieszczenia bloków układowych na chipie uzyskuje się dzieki wykorzystaniu algorytmu ewolucyjnego minimalizującego funkcję celu uwzględniającą poziom szumu podłożowego.
-
A Method of Real-Time Non-uniform Speech Stretching
PublicationDeveloped method of real-time non-uniform speech stretching is presented.The proposed solution is based on the well-known SOLA algorithm(Synchronous Overlap and Add). Non-uniform time-scale modification isachieved by the adjustment of time scaling factor values in accordance with thesignal content. Dependently on the speech unit (vowels/consonants), instantaneousrate of speech (ROS), and speech signal presence, values of the scalingfactor...
-
Self-Supervised Learning to Increase the Performance of Skin Lesion Classification
PublicationTo successfully train a deep neural network, a large amount of human-labeled data is required. Unfortunately, in many areas, collecting and labeling data is a difficult and tedious task. Several ways have been developed to mitigate the problem associated with the shortage of data, the most common of which is transfer learning. However, in many cases, the use of transfer learning as the only remedy is insufficient. In this study,...
-
Adaptive Hyperparameter Tuning within Neural Network-based Efficient Global Optimization
PublicationIn this paper, adaptive hyperparameter optimization (HPO) strategies within the efficient global optimization (EGO) with neural network (NN)-based prediction and uncertainty (EGONN) algorithm are proposed. These strategies utilize Bayesian optimization and multiarmed bandit optimization to tune HPs during the sequential sampling process either every iteration (HPO-1itr) or every five iterations (HPO-5itr). Through experiments using...
-
Frequently updated noise threat maps created with use of supercomputing grid
PublicationAn innovative supercomputing grid services devoted to noise threat evaluation were presented. The services described in this paper concern two issues, first is related to the noise mapping, while the second one focuses on assessment of the noise dose and its influence on the human hearing system. The discussed services were developed within the PL-Grid Plus Infrastructure which accumulates Polish academic supercomputer centers....
-
Time-domain prosodic modifications for text-to-speech synthesizer
PublicationAn application of prosodic speech processing algorithms to Text-To-Speech synthesis is presented. Prosodic modifications that improve the naturalness of the synthesized signal are discussed. The applied method is based on the TD-PSOLA algorithm. The developed Text-To-Speech Synthesizer is used in applications employing multimodal computer interfaces.
-
Abdominal Aortic Aneurysm segmentation from contrast-enhanced computed tomography angiography using deep convolutional networks
PublicationOne of the most common imaging methods for diagnosing an abdominal aortic aneurysm, and an endoleak detection is computed tomography angiography. In this paper, we address the problem of aorta and thrombus semantic segmentation, what is a mandatory step to estimate aortic aneurysm diameter. Three end-to-end convolutional neural networks were trained and evaluated. Finally, we proposed an ensemble of deep neural networks with underlying...
-
Application of Analytic Signal and Smooth Interpolation in Pulse Width Modulation for Conventional Matrix Converters
PublicationThe paper proposes an alternative and novel approach to the PWM duty cycles computation for Conventional Matrix Converters (CMC) fed by balanced, unbalanced or non–sinusoidal AC voltage sources. The presented solution simplifies the prototyping of direct modulation algorithms. PWM duty cycles are calculated faster by the smooth interpolation technique, using only vector coordinates, without trigonometric functions and angles. Both...
-
Artificial Neural Network for Multiprocessor Tasks Scheduling
Publication -
Approximation task decomposition for artificial neural network.
PublicationW pracy przedstawiono wpływ dekompozycji zadania na czasochłonność projektowania oraz dokładność i szybkość obliczeń sztucznej sieci neuronowej wykorzystanej do rozwiązania rzeczywistego problemu technicznego, którego matematyczny model był znany. Celem obliczeń prowadzonych przez sieć neuronową było określenie wartości współczynnika przepływu m na podstawie znajomości wartości: przewodności dźwiękowej C i średnicy przewodu d (a...
-
Towards classification of patients based on surface EMG data of temporomandibular joint muscles using self-organising maps
PublicationThe study considers the need for an effective method of classification of patients with a temporomandibular joint disorder (TMD). The self-organising map method (SOM) was applied to group patients and used together with the cross-correlation approach to interpret the processed (rectified and smoothed by using root mean square (RMS) algorithm) surface electromyography signal (sEMG) obtained from testing the muscles (two temporal...