The estimation of the frequency and decay factor of a single decaying exponential in noise is an important problem that has commanded significant attention in the research literature. In this paper, we examine the performance of computationally simple, yet accurate and robust, discrete Fourier transform based estimators. These estimators employ a coarse search followed by an interpolation step to obtain precise estimates of the parameters. Although their performance initially improves as the number of samples is increased, it reaches a minimum before departing significantly from the Cramer Rao Lower Bound (CRLB). We tackle this problem in this paper and propose a windowing strategy that allows the estimation performance to track the CRLB for a large number of samples. Furthermore, a practical implementation of the proposed method is given. Extensive simulations are reported that demonstrate the effectiveness of the proposed strategy.
{"title":"Windowed iterative estimation of the parameters of a damped complex exponential in noise","authors":"E. Aboutanios","doi":"10.5281/ZENODO.43590","DOIUrl":"https://doi.org/10.5281/ZENODO.43590","url":null,"abstract":"The estimation of the frequency and decay factor of a single decaying exponential in noise is an important problem that has commanded significant attention in the research literature. In this paper, we examine the performance of computationally simple, yet accurate and robust, discrete Fourier transform based estimators. These estimators employ a coarse search followed by an interpolation step to obtain precise estimates of the parameters. Although their performance initially improves as the number of samples is increased, it reaches a minimum before departing significantly from the Cramer Rao Lower Bound (CRLB). We tackle this problem in this paper and propose a windowing strategy that allows the estimation performance to track the CRLB for a large number of samples. Furthermore, a practical implementation of the proposed method is given. Extensive simulations are reported that demonstrate the effectiveness of the proposed strategy.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132429622","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we present a localised Capon spectral estimator that exhibits improved ability of resolving closely spaced peaks in the frequency domain. Starting with a unitary transformation using the discrete Fourier transform results in signals of interest being localised to small regions of the spectrum. Therefore, the algorithm uses this fact to construct a covariance matrix with low dimensionality, which can be easily inverted. This reduces the number of samples required to be averaged to produce the covariance matrix and as a result allows the spectrum resolution to be improved. In the simulation results, by comparing with the original Capon, the effectiveness of the novel estimator in resolving close peaks is demonstrated using both simulated undamped exponentials and simulated NMR spectroscopy signal (that is damped exponentials).
{"title":"Localised Capon spectral estimator with application to the processing of NMR signal","authors":"Shanglin Ye, E. Aboutanios","doi":"10.5281/ZENODO.43560","DOIUrl":"https://doi.org/10.5281/ZENODO.43560","url":null,"abstract":"In this paper, we present a localised Capon spectral estimator that exhibits improved ability of resolving closely spaced peaks in the frequency domain. Starting with a unitary transformation using the discrete Fourier transform results in signals of interest being localised to small regions of the spectrum. Therefore, the algorithm uses this fact to construct a covariance matrix with low dimensionality, which can be easily inverted. This reduces the number of samples required to be averaged to produce the covariance matrix and as a result allows the spectrum resolution to be improved. In the simulation results, by comparing with the original Capon, the effectiveness of the novel estimator in resolving close peaks is demonstrated using both simulated undamped exponentials and simulated NMR spectroscopy signal (that is damped exponentials).","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130939373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jean-Pierre Da Costa, Stefan Oprean, P. Baylou, C. Germain
The use of 3D imaging techniques is a choice approach for the study of the inner structure of materials. However, for large industrial applications, the stereological analysis of 2D snapshots of material sections is still necessary for obvious time and cost reasons. We present a novel method to analyze the 3D layout of cylindrical structures from a single 2D section. In particular we propose to estimate the distribution of cutting angles i.e. angles between the cylinders axis and the normal to the image plane. Contrary to existing approaches, the knowledge of the cylinder cross section shape is not a prerequisite. The only required input is the statistical distribution of the cylinder cross section area. Our approach is based on the minimization of a least squares criterion under linear constraints. It is evaluated on synthetic data and applied to microscopy images of fibrous composites. Our experimental study focuses on the capabilities and limitations of the approach.
{"title":"Estimating the distribution of 3D generalized cylinders angles from an image","authors":"Jean-Pierre Da Costa, Stefan Oprean, P. Baylou, C. Germain","doi":"10.5281/ZENODO.43326","DOIUrl":"https://doi.org/10.5281/ZENODO.43326","url":null,"abstract":"The use of 3D imaging techniques is a choice approach for the study of the inner structure of materials. However, for large industrial applications, the stereological analysis of 2D snapshots of material sections is still necessary for obvious time and cost reasons. We present a novel method to analyze the 3D layout of cylindrical structures from a single 2D section. In particular we propose to estimate the distribution of cutting angles i.e. angles between the cylinders axis and the normal to the image plane. Contrary to existing approaches, the knowledge of the cylinder cross section shape is not a prerequisite. The only required input is the statistical distribution of the cylinder cross section area. Our approach is based on the minimization of a least squares criterion under linear constraints. It is evaluated on synthetic data and applied to microscopy images of fibrous composites. Our experimental study focuses on the capabilities and limitations of the approach.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130217946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Energy efficiency emerges as a big concern in wireless sensor networks, where the lifetime of the system depends heavily on energy consumption. In a sensor network, the energy is consumed both by the circuit and the radio link. The radio link performance can be improved by using spatial multiplexing, but the conventional way of spatial multiplexing requires large antenna space and consumes more energy. Thus single RF MIMO transmission is proposed, which maps MIMO symbols on the radiation patterns to reduce power consumption. This paper presents the energy efficiency analysis over different schemes, which extends single RF MIMO capability with 16-QAM constellation though compact ESPAR antenna with only 2 elements. A detailed analysis on the minimal required energy is given, which shows that the proposed single RF MIMO with 16-QAM signaling has similar bit error rate performance with better energy efficiency.
{"title":"Energy efficient MIMO transmission with high order modulations for wireless sensor network","authors":"B. Han, A. Kalis, C. Papadias, R. Prasad","doi":"10.5281/ZENODO.43614","DOIUrl":"https://doi.org/10.5281/ZENODO.43614","url":null,"abstract":"Energy efficiency emerges as a big concern in wireless sensor networks, where the lifetime of the system depends heavily on energy consumption. In a sensor network, the energy is consumed both by the circuit and the radio link. The radio link performance can be improved by using spatial multiplexing, but the conventional way of spatial multiplexing requires large antenna space and consumes more energy. Thus single RF MIMO transmission is proposed, which maps MIMO symbols on the radiation patterns to reduce power consumption. This paper presents the energy efficiency analysis over different schemes, which extends single RF MIMO capability with 16-QAM constellation though compact ESPAR antenna with only 2 elements. A detailed analysis on the minimal required energy is given, which shows that the proposed single RF MIMO with 16-QAM signaling has similar bit error rate performance with better energy efficiency.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130482144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Subspace-based methods have been effectively used to estimate multi-input/multi-output, linear-time-invariant systems from noisy spectrum samples. In these methods, a critical step is splitting of two invariant subspaces associated with causal and non-causal eigenvalues of some structured matrices built from spectrum measurements via singular-value decomposition in order to determine model order. Mirror image symmetry with respect to the unit circle between the eigenvalue sets of the invariant spaces, required by these algorithms, is lost due to low signal-to-noise ratio, unmodelled dynamics, and insufficient amount of data. Consequently, the choice of model order is not straightforward. In this paper, we propose a robust model order selection scheme based on regularized nuclear norm optimization in combination with a recent subspace algorithm, which uses non-uniformly spaced, in frequencies, spectrum measurements. A simulation example shows the effectiveness of the proposed scheme to large amplitude noise over short data records. Then, the proposed scheme is used to design a linear-shape filter for random road excitations.
{"title":"Subspace-based spectrum estimation by nuclear norm minimization","authors":"H. Akçay, S. Türkay","doi":"10.5281/ZENODO.43436","DOIUrl":"https://doi.org/10.5281/ZENODO.43436","url":null,"abstract":"Subspace-based methods have been effectively used to estimate multi-input/multi-output, linear-time-invariant systems from noisy spectrum samples. In these methods, a critical step is splitting of two invariant subspaces associated with causal and non-causal eigenvalues of some structured matrices built from spectrum measurements via singular-value decomposition in order to determine model order. Mirror image symmetry with respect to the unit circle between the eigenvalue sets of the invariant spaces, required by these algorithms, is lost due to low signal-to-noise ratio, unmodelled dynamics, and insufficient amount of data. Consequently, the choice of model order is not straightforward. In this paper, we propose a robust model order selection scheme based on regularized nuclear norm optimization in combination with a recent subspace algorithm, which uses non-uniformly spaced, in frequencies, spectrum measurements. A simulation example shows the effectiveness of the proposed scheme to large amplitude noise over short data records. Then, the proposed scheme is used to design a linear-shape filter for random road excitations.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127873392","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Tracking multiple speakers with microphone arrays is used for practical applications such as video conferencing. An important task is the integration of multiple arrays with correct associations of multiple concurrent speakers. A single-array tracking approach based on CASA is extended here to probabilistic tracking with multiple arrays in order to handle a varying number of moving speakers over time and assign the concurrent localizations of multiple sensors to the speakers. Tracking is done simultaneously in angular and Euclidean space. The effectiveness of the method is shown with recordings of real speakers in a reverberant conference room by evaluation on the publicly available AV16.3 corpus.
{"title":"Online multi-speaker tracking using multiple microphone arrays informed by auditory scene analysis","authors":"A. Plinge, G. Fink","doi":"10.5281/ZENODO.43511","DOIUrl":"https://doi.org/10.5281/ZENODO.43511","url":null,"abstract":"Tracking multiple speakers with microphone arrays is used for practical applications such as video conferencing. An important task is the integration of multiple arrays with correct associations of multiple concurrent speakers. A single-array tracking approach based on CASA is extended here to probabilistic tracking with multiple arrays in order to handle a varying number of moving speakers over time and assign the concurrent localizations of multiple sensors to the speakers. Tracking is done simultaneously in angular and Euclidean space. The effectiveness of the method is shown with recordings of real speakers in a reverberant conference room by evaluation on the publicly available AV16.3 corpus.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"112 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114299103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper considers optimization and implementation of the pulse shape for Orthogonal Frequency Division Multiplexing (OFDM) systems over doubly dispersive channel. The proposed pulse shape is expressed as a linear combination of the most localized Hermite waveforms. In our previous works, we have proposed pulses optimization based on an exact expression of the Signal to Interference Ratio (SIR). In order to accelerate the optimization procedure, we propose an approximate expression of the SIR using Taylor series of interference and desired signal mean powers. For a performance evaluation of the optimized OFDM system, we propose an efficient implementation of both the modulator and the demodulator of the system and an efficient method to find out the resulting simulated SIR. Simulations results, obtained for different values of the dispersion factor, show that the SIR computed through system level simulation matches the maximum achievable SIR obtained through numerical optimization.
{"title":"Optimal OFDM pulse design, analysis and implementation over doubly dispersive channel","authors":"Raouia Ayadi, I. Kammoun, M. Siala","doi":"10.5281/ZENODO.43676","DOIUrl":"https://doi.org/10.5281/ZENODO.43676","url":null,"abstract":"This paper considers optimization and implementation of the pulse shape for Orthogonal Frequency Division Multiplexing (OFDM) systems over doubly dispersive channel. The proposed pulse shape is expressed as a linear combination of the most localized Hermite waveforms. In our previous works, we have proposed pulses optimization based on an exact expression of the Signal to Interference Ratio (SIR). In order to accelerate the optimization procedure, we propose an approximate expression of the SIR using Taylor series of interference and desired signal mean powers. For a performance evaluation of the optimized OFDM system, we propose an efficient implementation of both the modulator and the demodulator of the system and an efficient method to find out the resulting simulated SIR. Simulations results, obtained for different values of the dispersion factor, show that the SIR computed through system level simulation matches the maximum achievable SIR obtained through numerical optimization.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"6 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131487009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To identify biologically meaningful modules in large-scale biological networks, general blockmodel network clustering algorithms have recently attracted much attention to search for groups of molecules that have similar interaction patterns in networks. However, existing blockmodel module identification algorithms suffer from the problems of prohibitive computational complexity and being trapped at local optima due to its inherent combinatorial complexity. In this paper, we propose a novel blockmodel module identification formulation based on Markov random walk to address those problems by finding high quality approximate solutions. A new convex optimization problem is formulated to find the low conductance (LC) sets as potential modules based on the two-hop transition matrix of Markov random walk on networks. We further propose a spectral approximate algorithm to find high quality modules in large-scale networks. The experimental results on two real-world PPI (protein-protein interaction) networks demonstrate that our method outperforms the state-of-the-art blockmodel module identification algorithms in terms of the accuracy measured by the F-measure based on curated annotations such as GO (Gene Ontology) and KOG (EuKaryotic Orthologous Groups) categories.
{"title":"Blockmodel module identification in protein interaction networks through Markov random walk","authors":"Yijie Wang, Xiaoning Qian","doi":"10.5281/ZENODO.43423","DOIUrl":"https://doi.org/10.5281/ZENODO.43423","url":null,"abstract":"To identify biologically meaningful modules in large-scale biological networks, general blockmodel network clustering algorithms have recently attracted much attention to search for groups of molecules that have similar interaction patterns in networks. However, existing blockmodel module identification algorithms suffer from the problems of prohibitive computational complexity and being trapped at local optima due to its inherent combinatorial complexity. In this paper, we propose a novel blockmodel module identification formulation based on Markov random walk to address those problems by finding high quality approximate solutions. A new convex optimization problem is formulated to find the low conductance (LC) sets as potential modules based on the two-hop transition matrix of Markov random walk on networks. We further propose a spectral approximate algorithm to find high quality modules in large-scale networks. The experimental results on two real-world PPI (protein-protein interaction) networks demonstrate that our method outperforms the state-of-the-art blockmodel module identification algorithms in terms of the accuracy measured by the F-measure based on curated annotations such as GO (Gene Ontology) and KOG (EuKaryotic Orthologous Groups) categories.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134311808","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Segmentation of rib bones from magnetic resonance (MR) data is a challenging task due to the very weak MR signal of the bone tissue. The locations of the rib bones are indicated by the lack of signal between the surrounding soft tissues. We have developed an automated rib segmentation method, which tracks the bones from a given initial point using a predefined rib cross-section template. The three-dimensional location of the rib midline is parametrized with second order polynomials. The correct rib midline is determined through numerical optimization of polynomial parameters by maximizing cross-correlation between the template and the MR data. The segmentation result is a three-dimensional binary segmentation of the ribs, which can be output also as a surface. Comparison between the method's output and a manual segmentation shows that the method works adequately.
{"title":"Segmentation of rib bones from MR data using tracking approach and polynomial parametrization","authors":"Harri Pölönen, G. Schubert","doi":"10.5281/ZENODO.43439","DOIUrl":"https://doi.org/10.5281/ZENODO.43439","url":null,"abstract":"Segmentation of rib bones from magnetic resonance (MR) data is a challenging task due to the very weak MR signal of the bone tissue. The locations of the rib bones are indicated by the lack of signal between the surrounding soft tissues. We have developed an automated rib segmentation method, which tracks the bones from a given initial point using a predefined rib cross-section template. The three-dimensional location of the rib midline is parametrized with second order polynomials. The correct rib midline is determined through numerical optimization of polynomial parameters by maximizing cross-correlation between the template and the MR data. The segmentation result is a three-dimensional binary segmentation of the ribs, which can be output also as a surface. Comparison between the method's output and a manual segmentation shows that the method works adequately.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133233910","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In relation to speech enhancement, we study the influence of modifying the harmonic signal model for voiced speech to include small perturbations in the frequencies of the harmonics. A perturbed signal model is incorporated in the nonlinear least squares method, the Capon filter and the amplitude and phase estimation filter. Results show that it is possible to increase the performance, in terms of the signal reduction factor and the output signal-to-noise ratio, at the cost of increased complexity in the estimation of the model parameters. It is found that the perturbed signal model performs better than the harmonic signal model at input signal-to-noise ratios above approximately -10 dB, and that they are equally good below.
{"title":"On the influence of inharmonicities in model-based speech enhancement","authors":"S. Nørholm, J. Jensen, M. G. Christensen","doi":"10.5281/ZENODO.43372","DOIUrl":"https://doi.org/10.5281/ZENODO.43372","url":null,"abstract":"In relation to speech enhancement, we study the influence of modifying the harmonic signal model for voiced speech to include small perturbations in the frequencies of the harmonics. A perturbed signal model is incorporated in the nonlinear least squares method, the Capon filter and the amplitude and phase estimation filter. Results show that it is possible to increase the performance, in terms of the signal reduction factor and the output signal-to-noise ratio, at the cost of increased complexity in the estimation of the model parameters. It is found that the perturbed signal model performs better than the harmonic signal model at input signal-to-noise ratios above approximately -10 dB, and that they are equally good below.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133210441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：