2019 International Symposium on Signals, Circuits and Systems (ISSCS), 2019
In this paper we discuss how does the input magnitude data setting influence the behavior of erro... more In this paper we discuss how does the input magnitude data setting influence the behavior of error-reduction algorithm in the case of the one-dimensional discrete phase retrieval problem. We present experimental results related to the convergence or stagnation of the algorithm. We also discuss the issue of the zeros distribution of the solution, when the solution of the problem exists.
2015 International Symposium on Signals, Circuits and Systems (ISSCS), 2015
It has been known for decades that the iterative methods are perhaps the most popular approaches ... more It has been known for decades that the iterative methods are perhaps the most popular approaches to solve the phase retrieval problem. Unfortunately the iterative methods often stagnate. This happens also in the case of the 1-D Discrete Phase Retrieval (1-D DPhR) problem. Recently it has been shown that some requirements in the input magnitude data might be one of the reasons why the direct method cannot solve the 1-D DPhR Problem. In this work we present some difficulties that can be encountered when one has to implement the iterative method for finding a solution of 1-D DPhR problem. We shall formulate the extended form of 1-D DPhR problem. Simulations indicate the conjecture to be true.
The goal of this paper is to detail the Analog Filter Netlist to Digital Filter Statements approa... more The goal of this paper is to detail the Analog Filter Netlist to Digital Filter Statements approach used to obtain the digital filter from the netlist of the analog filter. This method is done by replacing each component by its corresponding companion model and then discretizing the set of equations which describe in the time domain the behavior of all the circuit’s components. Finally, the software implementation of the digital filter consists in the transcription of the discrete-time system of equations. We show that the selection of step-size might be sensitive problem for this approach. Keywords—analog filter; digital filter; netlist
Carpathian Journal of Electronic and Computer Engineering, 2021
In this paper we present an updated version of the audio database acquired by the TIAGo service r... more In this paper we present an updated version of the audio database acquired by the TIAGo service robot and by the simulated TIAGo service robot. To the initial database which consists in 1380 audio signals we have added 1920 more acoustic signals. The audio database consists now in 3300 isolated audio events corresponding to 110 classes. All the recorded sound events correspond to the indoor environment, and they are spread into five different scenarios: kitchen, room, appliances, voice and non-verbal. The audio database is intended to be used in order to identify indoor events based on audio signature, especially when elderly or chronically ill people live alone.
2020 28th European Signal Processing Conference (EUSIPCO), 2021
Many signals in nature and engineering systems can be locally modeled as relatively low degree po... more Many signals in nature and engineering systems can be locally modeled as relatively low degree polynomials, thus one-dimensional polynomial predictive filters are useful especially in time-critical systems. The goal of this paper is to introduce the two-dimensional polynomial predictive FIR filters and present few of their properties. First we discuss previous main results in one-dimensional polynomial predictive filters. Then we show how to find the coefficients and the system functions of the minimum area polynomial predictor, and we present the recursive form for the system function of a minimum area polynomial predictor. Finally, we approach the general form of 2D polynomial predictors.
Electrical activity of neuronal cells is thought to support brain function, being modified in dis... more Electrical activity of neuronal cells is thought to support brain function, being modified in disease compared to health conditions. The tools for analyzing such data are diverse and need to be tuned for each particular application. We propose a wide and diverse perspective characterization of induced epileptic activity in zebrafish field recordings. We have found that features regarding amplitude, width, inter-event-intervals and autocorrelation profiles can be automatically extracted and used as markers to differentiate between different genetic or pharmacologic conditions. The present study aims at developing a toolbox that aids the understanding of the underlying mechanisms of epileptiform discharges. Keywords— electrophysiology, epilepsy, signal processing, ictal, neural;
2017 14th International Conference on Engineering of Modern Electric Systems (EMES), 2017
This paper presents a grid search approach to optimize the kernel's parameters for the suppor... more This paper presents a grid search approach to optimize the kernel's parameters for the support vector machines classifier. The most encountered three kernels are considered: linear, radial basis, and sigmoid. We show that the optimization of parameters improves the recognition performance for audio signals classification, especially in the case of sigmoid kernel. The behavior of the model is very sensitive to kernel's parameters, which, in turn are sensitive to data and selected features. We consider the problem of multiclass classification with imbalanced datasets. We compare the accuracies obtain with and without kernel parameter's optimization. As features we use Mel frequency cepstral coefficients.
2017 International Conference on Speech Technology and Human-Computer Dialogue (SpeD), 2017
The goal of this work is to present some possible intruder detection systems and the influence of... more The goal of this work is to present some possible intruder detection systems and the influence of impulse-like signals upon the overall classification accuracy. Two different scenarios are used: in the first scenario five sound classes are considered (last class belong to impulsive sounds — gunshots), while in the second scenario we dropped out the impulsive sound class. More classifiers are used in both scenarios and different number of features are considered. An improvement in the classification accuracy is obtained within the second scenario. The highest accuracy for the first scenario is for J48 classifier using 51 features, while for the second scenario the highest accuracy is attained for Simple Logistic classifier wit 101 features.
2020 13th International Conference on Communications (COMM), 2020
The task of creating a short, accurate and fluent summary starting from a larger text document or... more The task of creating a short, accurate and fluent summary starting from a larger text document or group of documents is called text summarization. When the summary is generated by extracting parts from the original text, the process is known as extractive summarization. This work is focused on the use of convex optimization and positive defined symmetric kernels for the extractive summarization of a text. The paper includes two new contributions. First, we show how the kernel method can be used for unsupervised extractive summarization. Second, we investigate empirically the performance of different kernel functions with respect to quality and execution time.
2016 24th European Signal Processing Conference (EUSIPCO), 2016
In this paper some results on Schur transform are reviewed to address the problem of one-dimensio... more In this paper some results on Schur transform are reviewed to address the problem of one-dimensional discrete phase retrieval. The goal is to provide a test whether a sequence of input magnitude data gives a solution to one-dimensional discrete phase retrieval problem. It has been previously shown that this issue is related to the nonnegativity of trigonometric polynomials. The proposed method is similar to the table procedure for counting the multiplicities of zeros on unit circle. Examples and numerical results are also provided to indicate that the problem of one-dimensional discrete phase retrieval often does not have a solution.
We consider the discrete form of the one-dimensional phase retrieval (1-D DPhR) problem from the ... more We consider the discrete form of the one-dimensional phase retrieval (1-D DPhR) problem from the point of view of input magnitude data. The direct method can provide a solution to the 1-D DPhR problem if certain conditions are satisfied by the input magnitude data, namely the corresponding trigonometric polynomial must be nonnegative. To test positivity of a trigonometric polynomial a novel DFT-based criterion is proposed. We use this DFT criterion for different sets of input magnitude data to evaluate whether the direct method applied to the 1-D DPhR problem leads to a solution in all explored cases.
In this paper we present the influence of quantization of audio signals on D and S descriptors. T... more In this paper we present the influence of quantization of audio signals on D and S descriptors. These descriptors represent the number of samples between two consecutive real zeros and the number of points of local minima/maxima between two successive real zeros, respectively. It is shown that the number of the D/S pairs is almost constant and the behavior is almost the same till the number of bits is less than 6, in the proposed audio based wildlife intruder detection framework.
The resolution of ultrasound medical images is yet an important problem despite of the researcher... more The resolution of ultrasound medical images is yet an important problem despite of the researchers efforts. In this paper we presents a nonlinear blind deconvolution to eliminate the blurring effect based on the measured radiofrequency signal envelope. This algorithm is executed in two steps. Firslty we make an estimation for Point Spread Function (PSF) and, secondly we use the estimated PSF to remove, iteratively their effect. The proposed algorithm is a greedy algorithm, called also matching pursuit or CLEAN. The use of this algorithm is motivated beacause theorically it avoid the so called inverse problem, which usually needs regularization to obtain an optimal solution. The results are presented using 1D simulated signals in term of visual evaluation and nMSE in comparison with the two most kwown regularisation solution methods for least square problem, Thikonov regularization or l2-norm and Total Variation or l1 norm.
After a review of the circle fitting issue, we recall a relatively unknown method derived from a ... more After a review of the circle fitting issue, we recall a relatively unknown method derived from a classical geometric result. We propose an improvement of this technique by reweighting the data, iterating the procedure, and choosing at every step as the new inversion point the one diametrically opposite to the previous inversion point.
2018 26th European Signal Processing Conference (EUSIPCO)
Recently it has been shown that the one-dimensional discrete phase retrieval problem may not have... more Recently it has been shown that the one-dimensional discrete phase retrieval problem may not have always a causal solution for certain input magnitude data, but it has been proved that the extended form of the one-dimensional discrete phase retrieval problem has always a causal solution within the same conditions. In this work we are looking for the minimum length solution for one-dimensional discrete phase retrieval problem. The Non-uniform Discrete Fourier Transform based approach is introduced and experimental results are also presented.
2019 International Symposium on Signals, Circuits and Systems (ISSCS), 2019
In this paper we discuss how does the input magnitude data setting influence the behavior of erro... more In this paper we discuss how does the input magnitude data setting influence the behavior of error-reduction algorithm in the case of the one-dimensional discrete phase retrieval problem. We present experimental results related to the convergence or stagnation of the algorithm. We also discuss the issue of the zeros distribution of the solution, when the solution of the problem exists.
2015 International Symposium on Signals, Circuits and Systems (ISSCS), 2015
It has been known for decades that the iterative methods are perhaps the most popular approaches ... more It has been known for decades that the iterative methods are perhaps the most popular approaches to solve the phase retrieval problem. Unfortunately the iterative methods often stagnate. This happens also in the case of the 1-D Discrete Phase Retrieval (1-D DPhR) problem. Recently it has been shown that some requirements in the input magnitude data might be one of the reasons why the direct method cannot solve the 1-D DPhR Problem. In this work we present some difficulties that can be encountered when one has to implement the iterative method for finding a solution of 1-D DPhR problem. We shall formulate the extended form of 1-D DPhR problem. Simulations indicate the conjecture to be true.
The goal of this paper is to detail the Analog Filter Netlist to Digital Filter Statements approa... more The goal of this paper is to detail the Analog Filter Netlist to Digital Filter Statements approach used to obtain the digital filter from the netlist of the analog filter. This method is done by replacing each component by its corresponding companion model and then discretizing the set of equations which describe in the time domain the behavior of all the circuit’s components. Finally, the software implementation of the digital filter consists in the transcription of the discrete-time system of equations. We show that the selection of step-size might be sensitive problem for this approach. Keywords—analog filter; digital filter; netlist
Carpathian Journal of Electronic and Computer Engineering, 2021
In this paper we present an updated version of the audio database acquired by the TIAGo service r... more In this paper we present an updated version of the audio database acquired by the TIAGo service robot and by the simulated TIAGo service robot. To the initial database which consists in 1380 audio signals we have added 1920 more acoustic signals. The audio database consists now in 3300 isolated audio events corresponding to 110 classes. All the recorded sound events correspond to the indoor environment, and they are spread into five different scenarios: kitchen, room, appliances, voice and non-verbal. The audio database is intended to be used in order to identify indoor events based on audio signature, especially when elderly or chronically ill people live alone.
2020 28th European Signal Processing Conference (EUSIPCO), 2021
Many signals in nature and engineering systems can be locally modeled as relatively low degree po... more Many signals in nature and engineering systems can be locally modeled as relatively low degree polynomials, thus one-dimensional polynomial predictive filters are useful especially in time-critical systems. The goal of this paper is to introduce the two-dimensional polynomial predictive FIR filters and present few of their properties. First we discuss previous main results in one-dimensional polynomial predictive filters. Then we show how to find the coefficients and the system functions of the minimum area polynomial predictor, and we present the recursive form for the system function of a minimum area polynomial predictor. Finally, we approach the general form of 2D polynomial predictors.
Electrical activity of neuronal cells is thought to support brain function, being modified in dis... more Electrical activity of neuronal cells is thought to support brain function, being modified in disease compared to health conditions. The tools for analyzing such data are diverse and need to be tuned for each particular application. We propose a wide and diverse perspective characterization of induced epileptic activity in zebrafish field recordings. We have found that features regarding amplitude, width, inter-event-intervals and autocorrelation profiles can be automatically extracted and used as markers to differentiate between different genetic or pharmacologic conditions. The present study aims at developing a toolbox that aids the understanding of the underlying mechanisms of epileptiform discharges. Keywords— electrophysiology, epilepsy, signal processing, ictal, neural;
2017 14th International Conference on Engineering of Modern Electric Systems (EMES), 2017
This paper presents a grid search approach to optimize the kernel's parameters for the suppor... more This paper presents a grid search approach to optimize the kernel's parameters for the support vector machines classifier. The most encountered three kernels are considered: linear, radial basis, and sigmoid. We show that the optimization of parameters improves the recognition performance for audio signals classification, especially in the case of sigmoid kernel. The behavior of the model is very sensitive to kernel's parameters, which, in turn are sensitive to data and selected features. We consider the problem of multiclass classification with imbalanced datasets. We compare the accuracies obtain with and without kernel parameter's optimization. As features we use Mel frequency cepstral coefficients.
2017 International Conference on Speech Technology and Human-Computer Dialogue (SpeD), 2017
The goal of this work is to present some possible intruder detection systems and the influence of... more The goal of this work is to present some possible intruder detection systems and the influence of impulse-like signals upon the overall classification accuracy. Two different scenarios are used: in the first scenario five sound classes are considered (last class belong to impulsive sounds — gunshots), while in the second scenario we dropped out the impulsive sound class. More classifiers are used in both scenarios and different number of features are considered. An improvement in the classification accuracy is obtained within the second scenario. The highest accuracy for the first scenario is for J48 classifier using 51 features, while for the second scenario the highest accuracy is attained for Simple Logistic classifier wit 101 features.
2020 13th International Conference on Communications (COMM), 2020
The task of creating a short, accurate and fluent summary starting from a larger text document or... more The task of creating a short, accurate and fluent summary starting from a larger text document or group of documents is called text summarization. When the summary is generated by extracting parts from the original text, the process is known as extractive summarization. This work is focused on the use of convex optimization and positive defined symmetric kernels for the extractive summarization of a text. The paper includes two new contributions. First, we show how the kernel method can be used for unsupervised extractive summarization. Second, we investigate empirically the performance of different kernel functions with respect to quality and execution time.
2016 24th European Signal Processing Conference (EUSIPCO), 2016
In this paper some results on Schur transform are reviewed to address the problem of one-dimensio... more In this paper some results on Schur transform are reviewed to address the problem of one-dimensional discrete phase retrieval. The goal is to provide a test whether a sequence of input magnitude data gives a solution to one-dimensional discrete phase retrieval problem. It has been previously shown that this issue is related to the nonnegativity of trigonometric polynomials. The proposed method is similar to the table procedure for counting the multiplicities of zeros on unit circle. Examples and numerical results are also provided to indicate that the problem of one-dimensional discrete phase retrieval often does not have a solution.
We consider the discrete form of the one-dimensional phase retrieval (1-D DPhR) problem from the ... more We consider the discrete form of the one-dimensional phase retrieval (1-D DPhR) problem from the point of view of input magnitude data. The direct method can provide a solution to the 1-D DPhR problem if certain conditions are satisfied by the input magnitude data, namely the corresponding trigonometric polynomial must be nonnegative. To test positivity of a trigonometric polynomial a novel DFT-based criterion is proposed. We use this DFT criterion for different sets of input magnitude data to evaluate whether the direct method applied to the 1-D DPhR problem leads to a solution in all explored cases.
In this paper we present the influence of quantization of audio signals on D and S descriptors. T... more In this paper we present the influence of quantization of audio signals on D and S descriptors. These descriptors represent the number of samples between two consecutive real zeros and the number of points of local minima/maxima between two successive real zeros, respectively. It is shown that the number of the D/S pairs is almost constant and the behavior is almost the same till the number of bits is less than 6, in the proposed audio based wildlife intruder detection framework.
The resolution of ultrasound medical images is yet an important problem despite of the researcher... more The resolution of ultrasound medical images is yet an important problem despite of the researchers efforts. In this paper we presents a nonlinear blind deconvolution to eliminate the blurring effect based on the measured radiofrequency signal envelope. This algorithm is executed in two steps. Firslty we make an estimation for Point Spread Function (PSF) and, secondly we use the estimated PSF to remove, iteratively their effect. The proposed algorithm is a greedy algorithm, called also matching pursuit or CLEAN. The use of this algorithm is motivated beacause theorically it avoid the so called inverse problem, which usually needs regularization to obtain an optimal solution. The results are presented using 1D simulated signals in term of visual evaluation and nMSE in comparison with the two most kwown regularisation solution methods for least square problem, Thikonov regularization or l2-norm and Total Variation or l1 norm.
After a review of the circle fitting issue, we recall a relatively unknown method derived from a ... more After a review of the circle fitting issue, we recall a relatively unknown method derived from a classical geometric result. We propose an improvement of this technique by reweighting the data, iterating the procedure, and choosing at every step as the new inversion point the one diametrically opposite to the previous inversion point.
2018 26th European Signal Processing Conference (EUSIPCO)
Recently it has been shown that the one-dimensional discrete phase retrieval problem may not have... more Recently it has been shown that the one-dimensional discrete phase retrieval problem may not have always a causal solution for certain input magnitude data, but it has been proved that the extended form of the one-dimensional discrete phase retrieval problem has always a causal solution within the same conditions. In this work we are looking for the minimum length solution for one-dimensional discrete phase retrieval problem. The Non-uniform Discrete Fourier Transform based approach is introduced and experimental results are also presented.
Uploads
Papers by Corneliu Rusu