Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613534
Z. Malouche, O. Macchi
The problem of separation of mixed sources is addressed in this paper. To solve this problem, at least as many observations as sources are needed. In particular, the number of sources can be unknown. The separation system is a linear network updated with a stochastic descent algorithm to minimize some separation criterion. A first algorithm separates sources with positive kurtosises while a second one separates sources with negative kurtosises. For both, the performances are independent of the mixture. Besides, in the noisy case, when there are more sensors than sources, the additional outputs merely generate noise.
{"title":"Adaptive separation of an unknown number of sources","authors":"Z. Malouche, O. Macchi","doi":"10.1109/HOST.1997.613534","DOIUrl":"https://doi.org/10.1109/HOST.1997.613534","url":null,"abstract":"The problem of separation of mixed sources is addressed in this paper. To solve this problem, at least as many observations as sources are needed. In particular, the number of sources can be unknown. The separation system is a linear network updated with a stochastic descent algorithm to minimize some separation criterion. A first algorithm separates sources with positive kurtosises while a second one separates sources with negative kurtosises. For both, the performances are independent of the mixture. Besides, in the noisy case, when there are more sensors than sources, the additional outputs merely generate noise.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133795606","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613560
D. Declercq, P. Duvant
This paper presents some asymptotical results of the Hermite normality test previously introduced. We show that the Hermite statistic S/sub H/ is distributed under the null hypothesis as a quadratic form of normal variates and under the nonnull hypothesis as normal. The special case of tests with two polynomials is studied in detail. Finally, we give some considerations for the choice of the best Hermite test when prior knowledge is available and especially we determine the test asymptotically the most powerful for a fixed alternative distribution (the uniform distribution). Those results are supported by simulations.
{"title":"Asymptotic distribution of the Hermite normality test","authors":"D. Declercq, P. Duvant","doi":"10.1109/HOST.1997.613560","DOIUrl":"https://doi.org/10.1109/HOST.1997.613560","url":null,"abstract":"This paper presents some asymptotical results of the Hermite normality test previously introduced. We show that the Hermite statistic S/sub H/ is distributed under the null hypothesis as a quadratic form of normal variates and under the nonnull hypothesis as normal. The special case of tests with two polynomials is studied in detail. Finally, we give some considerations for the choice of the best Hermite test when prior knowledge is available and especially we determine the test asymptotically the most powerful for a fixed alternative distribution (the uniform distribution). Those results are supported by simulations.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130096534","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613515
Chih-Chung Feng, Chong-Yung Chi
A unified class of inverse filter criteria using two cumulants, which includes Wiggins' criterion, Donoho's criteria and Tugnait's criteria as special cases, has been proposed by Chi and Wu for blind deconvolution and equalization of real single-input single-output (SISO) linear time-invariant (LTI) systems. In this paper, we extend this class of (single channel) inverse filter criteria to a family of multistage and a family of single stage criteria for deconvolution and equalization of real (or complex) multi-input multi-output (MIMO) LTI systems with only non-Gaussian measurements. It can be shown that the two families of inverse filter criteria lead to perfect equalization for MIMO systems under some conditions. Some simulation results for the optimum inverse filter using gradient type iterative optimization algorithms were provided to support the proposed criteria. Finally, we draw some conclusions.
{"title":"A class of blind deconvolution and equalization algorithms for nonminimum phase multi-input multi-output systems","authors":"Chih-Chung Feng, Chong-Yung Chi","doi":"10.1109/HOST.1997.613515","DOIUrl":"https://doi.org/10.1109/HOST.1997.613515","url":null,"abstract":"A unified class of inverse filter criteria using two cumulants, which includes Wiggins' criterion, Donoho's criteria and Tugnait's criteria as special cases, has been proposed by Chi and Wu for blind deconvolution and equalization of real single-input single-output (SISO) linear time-invariant (LTI) systems. In this paper, we extend this class of (single channel) inverse filter criteria to a family of multistage and a family of single stage criteria for deconvolution and equalization of real (or complex) multi-input multi-output (MIMO) LTI systems with only non-Gaussian measurements. It can be shown that the two families of inverse filter criteria lead to perfect equalization for MIMO systems under some conditions. Some simulation results for the optimum inverse filter using gradient type iterative optimization algorithms were provided to support the proposed criteria. Finally, we draw some conclusions.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130304046","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613549
E. Moreau
We consider the problem of sources separation and particularly criteria based approaches. A generalized definition of contrast function is given in order to consider non symmetrical and/or non scale invariant functions. Two generalized contrasts involving high-order cumulants are proposed. In the case of two sources, we derive the optimal non symmetrical coefficient by minimizing a performance index. Finally computer simulations are presented in order to illustrate the results and to show the interest in considering a non symmetrical contrast.
{"title":"Towards non symmetrical optimal source separation contrasts","authors":"E. Moreau","doi":"10.1109/HOST.1997.613549","DOIUrl":"https://doi.org/10.1109/HOST.1997.613549","url":null,"abstract":"We consider the problem of sources separation and particularly criteria based approaches. A generalized definition of contrast function is given in order to consider non symmetrical and/or non scale invariant functions. Two generalized contrasts involving high-order cumulants are proposed. In the case of two sources, we derive the optimal non symmetrical coefficient by minimizing a performance index. Finally computer simulations are presented in order to illustrate the results and to show the interest in considering a non symmetrical contrast.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127728395","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613562
G. Giunta, G. Jacovitti, G. Scarano
In this contribution we develop a procedure for deciding whether a finite segment of a signal can be considered as a realization of a Gaussian process or not. We first present the theoretical bases leading to the formulation of the Bussgang test. A novel test based on the sign nonlinearity is introduced and its performance analysed in comparison with two simple Gaussianity tests presented in the literature. The results have shown a very promising behaviour of the suggested method in the presence of small amount of available data in both the cases of white and correlated samples.
{"title":"Bussgang Gaussianity test for stationary series","authors":"G. Giunta, G. Jacovitti, G. Scarano","doi":"10.1109/HOST.1997.613562","DOIUrl":"https://doi.org/10.1109/HOST.1997.613562","url":null,"abstract":"In this contribution we develop a procedure for deciding whether a finite segment of a signal can be considered as a realization of a Gaussian process or not. We first present the theoretical bases leading to the formulation of the Bussgang test. A novel test based on the sign nonlinearity is introduced and its performance analysed in comparison with two simple Gaussianity tests presented in the literature. The results have shown a very promising behaviour of the suggested method in the presence of small amount of available data in both the cases of white and correlated samples.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129447105","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613488
D. Boss, T. Petermann, K. Kammeyer
We investigate in this paper whether the HOS-based blind channel estimation method EVI (eigenvector approach to blind identification) can compete with the nonblind cross-correlation-based scheme used in state-of-the-art GSM receivers (Global System for Mobile communication). For blind, non-blind, and ideal estimates of COST-207 mobile radio channels, we give simulated bit error rates (BER) after Viterbi detection in terms of the mean signal-to-noise ratio (SNR). Averaged over three COST-207 propagation environments, EVI leads to a mean SNR loss of 1.2 dB only, while it saves the 22% overhead in the GSM data rate due to the transmission of training sequences. Since just 142 samples are used for channel estimation, we consider this performance outstanding for an approach based on HOS.
{"title":"Impact of blind versus non-blind channel estimation on the BER performance of GSM receivers","authors":"D. Boss, T. Petermann, K. Kammeyer","doi":"10.1109/HOST.1997.613488","DOIUrl":"https://doi.org/10.1109/HOST.1997.613488","url":null,"abstract":"We investigate in this paper whether the HOS-based blind channel estimation method EVI (eigenvector approach to blind identification) can compete with the nonblind cross-correlation-based scheme used in state-of-the-art GSM receivers (Global System for Mobile communication). For blind, non-blind, and ideal estimates of COST-207 mobile radio channels, we give simulated bit error rates (BER) after Viterbi detection in terms of the mean signal-to-noise ratio (SNR). Averaged over three COST-207 propagation environments, EVI leads to a mean SNR loss of 1.2 dB only, while it saves the 22% overhead in the GSM data rate due to the transmission of training sequences. Since just 142 samples are used for channel estimation, we consider this performance outstanding for an approach based on HOS.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121463742","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613486
K. J. Turner, F. Faruqi, C.L. Brown
A sub-optimal nonlinear time-recursive filter is developed which considers an arbitrary number of moments of the conditional density. The filter assumes a quadratic truncation of the system dynamics and measurement functions and retains N moments, thus requiring knowledge of up to 2N+2 a priori moments and N+1 moments of the measurement noise process, which may be non-Gaussian. Prediction and update relations are given for moments of arbitrary order along with mechanisms which facilitate their closed forms. Numerical examples are given for both scalar and vector systems and show promising results.
{"title":"Tracking nonlinear systems using higher order moments","authors":"K. J. Turner, F. Faruqi, C.L. Brown","doi":"10.1109/HOST.1997.613486","DOIUrl":"https://doi.org/10.1109/HOST.1997.613486","url":null,"abstract":"A sub-optimal nonlinear time-recursive filter is developed which considers an arbitrary number of moments of the conditional density. The filter assumes a quadratic truncation of the system dynamics and measurement functions and retains N moments, thus requiring knowledge of up to 2N+2 a priori moments and N+1 moments of the measurement noise process, which may be non-Gaussian. Prediction and update relations are given for moments of arbitrary order along with mechanisms which facilitate their closed forms. Numerical examples are given for both scalar and vector systems and show promising results.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120946544","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613513
C. Huet, J. Le Roux
In this paper we give an analytic optimal solution to the identification problem of non minimum phase systems using the fourth order spectra. We show that this solution is in first approximation equivalent to the solution given by the well-known kurtosis maximization method. The proposed solution gives the phase of the system transfer function, the modulus can be obtained from the second order statistics. However this solution requires trispectrum phase unwrapping as the trispectrum phase is known in the interval [-/spl pi/,/spl pi/] but needs to be unwrapped in the interval [-4/spl pi/,4/spl pi/] in order to obtain the optimal solution. Therefore, we present different phase unwrapping solutions. Next, we propose a method to improve the trispectrum phase estimation using a factorizability condition. Simulation results are given and the algorithm shows good behavior even with few data.
{"title":"Blind system identification using fourth order spectral analysis of complex signals","authors":"C. Huet, J. Le Roux","doi":"10.1109/HOST.1997.613513","DOIUrl":"https://doi.org/10.1109/HOST.1997.613513","url":null,"abstract":"In this paper we give an analytic optimal solution to the identification problem of non minimum phase systems using the fourth order spectra. We show that this solution is in first approximation equivalent to the solution given by the well-known kurtosis maximization method. The proposed solution gives the phase of the system transfer function, the modulus can be obtained from the second order statistics. However this solution requires trispectrum phase unwrapping as the trispectrum phase is known in the interval [-/spl pi/,/spl pi/] but needs to be unwrapped in the interval [-4/spl pi/,4/spl pi/] in order to obtain the optimal solution. Therefore, we present different phase unwrapping solutions. Next, we propose a method to improve the trispectrum phase estimation using a factorizability condition. Simulation results are given and the algorithm shows good behavior even with few data.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123085784","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613556
L. A. Pflug, G. Ioup, J. Ioup
A comparison between second and fourth order moment detectors is made for three low frequency transient signals embedded in measured ambient shipping noise from a shallow-water area. Detector evaluations must be made over short periods of time because of the nonstationary nature of the noise. Results indicate that the fourth order moment detects better than the second order moment for the two nonGaussian transient signals, but not as well for the more Gaussian transient signal. This suggests that the fourth order moment detector be used in addition to the second order moment detector, rather than as a replacement. The results of detection simulations with independent Gaussian noise are given for comparison to the results with the measured shipping noise. The comparison indicates that gains for the fourth order moment detector over the second order moment detector are higher in the shipping noise than in the Gaussian noise for the two nonGaussian signals.
{"title":"Higher order moment detection of transients in measured shallow-water noise","authors":"L. A. Pflug, G. Ioup, J. Ioup","doi":"10.1109/HOST.1997.613556","DOIUrl":"https://doi.org/10.1109/HOST.1997.613556","url":null,"abstract":"A comparison between second and fourth order moment detectors is made for three low frequency transient signals embedded in measured ambient shipping noise from a shallow-water area. Detector evaluations must be made over short periods of time because of the nonstationary nature of the noise. Results indicate that the fourth order moment detects better than the second order moment for the two nonGaussian transient signals, but not as well for the more Gaussian transient signal. This suggests that the fourth order moment detector be used in addition to the second order moment detector, rather than as a replacement. The results of detection simulations with independent Gaussian noise are given for comparison to the results with the measured shipping noise. The comparison indicates that gains for the fourth order moment detector over the second order moment detector are higher in the shipping noise than in the Gaussian noise for the two nonGaussian signals.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115836589","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}
Pub Date : 1997-07-21DOI: 10.1109/HOST.1997.613505
G. Krieger, C. Zetzsche, E. Barth
Natural images contain considerable statistical redundancies beyond the level of second-order correlations. To identify the nature of these higher-order dependencies, we analyze the bispectra and trispectra of natural images. Our investigations reveal substantial statistical dependencies between those frequency components which are aligned to each other with respect to orientation. We argue that operators which are selective to local intrinsic dimensionality can optimally exploit such redundancies. We also show that the polyspectral structure we find for natural images helps to understand the hitherto unexplained superiority of orientation-selective filter decompositions over isotropic schemes like the Laplacian pyramid. However any essentially linear scheme can only partially exploit this higher-order redundancy. We therefore propose nonlinear i2D-selective operators which exhibit close resemblance to hypercomplex and end-stopped cells in the visual cortex. The function of these operators can be interpreted as a higher-order whitening of the input signal.
{"title":"Higher-order statistics of natural images and their exploitation by operators selective to intrinsic dimensionality","authors":"G. Krieger, C. Zetzsche, E. Barth","doi":"10.1109/HOST.1997.613505","DOIUrl":"https://doi.org/10.1109/HOST.1997.613505","url":null,"abstract":"Natural images contain considerable statistical redundancies beyond the level of second-order correlations. To identify the nature of these higher-order dependencies, we analyze the bispectra and trispectra of natural images. Our investigations reveal substantial statistical dependencies between those frequency components which are aligned to each other with respect to orientation. We argue that operators which are selective to local intrinsic dimensionality can optimally exploit such redundancies. We also show that the polyspectral structure we find for natural images helps to understand the hitherto unexplained superiority of orientation-selective filter decompositions over isotropic schemes like the Laplacian pyramid. However any essentially linear scheme can only partially exploit this higher-order redundancy. We therefore propose nonlinear i2D-selective operators which exhibit close resemblance to hypercomplex and end-stopped cells in the visual cortex. The function of these operators can be interpreted as a higher-order whitening of the input signal.","PeriodicalId":305928,"journal":{"name":"Proceedings of the IEEE Signal Processing Workshop on Higher-Order Statistics","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131464689","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：