Alice Combernoux, Frédéric Pascal, G. Ginolhac, M. Lesturgie
The paper addresses the problem of target detection embedded in a disturbance composed of a low rank Gaussian clutter and a white Gaussian noise. In this context, it is interesting to use an adaptive version of the Low Rank Normalized Matched Filter detector, denoted LR-ANMF, which is a function of the estimation of the projector onto the clutter subspace. In this paper, we show that the LR-ANMF detector based on the sample covariance matrix is consistent when the number of secondary data K tends to infinity for a fixed data dimension m but not consistent when m and K both tend to infinity at the same rate, i.e. m/K → c ∈ (0, ∞). Using the results of random matrix theory, we then propose a new version of the LR-ANMF which is consistent in both cases. The application of our new detector on STAP (Space Time Adaptive Processing) data shows the interest of our approach.
{"title":"Random matrix theory applied to low rank stap detection","authors":"Alice Combernoux, Frédéric Pascal, G. Ginolhac, M. Lesturgie","doi":"10.5281/ZENODO.43697","DOIUrl":"https://doi.org/10.5281/ZENODO.43697","url":null,"abstract":"The paper addresses the problem of target detection embedded in a disturbance composed of a low rank Gaussian clutter and a white Gaussian noise. In this context, it is interesting to use an adaptive version of the Low Rank Normalized Matched Filter detector, denoted LR-ANMF, which is a function of the estimation of the projector onto the clutter subspace. In this paper, we show that the LR-ANMF detector based on the sample covariance matrix is consistent when the number of secondary data K tends to infinity for a fixed data dimension m but not consistent when m and K both tend to infinity at the same rate, i.e. m/K → c ∈ (0, ∞). Using the results of random matrix theory, we then propose a new version of the LR-ANMF which is consistent in both cases. The application of our new detector on STAP (Space Time Adaptive Processing) data shows the interest of our approach.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"34 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":"128881964","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, the kernel proportionate normalized least mean square algorithm (KPNLMS) is proposed. The proportionate factors are used in order to increase the convergence speed and the tracking abilities of the kernel normalized least mean square (KNLMS) adaptive algorithm. We confirm the effectiveness of the proposed algorithm for nonlinear system identification and forward prediction using computer simulations.
{"title":"The kernel proportionate NLMS algorithm","authors":"F. Albu, K. Nishikawa","doi":"10.5281/ZENODO.43315","DOIUrl":"https://doi.org/10.5281/ZENODO.43315","url":null,"abstract":"In this paper, the kernel proportionate normalized least mean square algorithm (KPNLMS) is proposed. The proportionate factors are used in order to increase the convergence speed and the tracking abilities of the kernel normalized least mean square (KNLMS) adaptive algorithm. We confirm the effectiveness of the proposed algorithm for nonlinear system identification and forward prediction using computer simulations.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"337 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":"115670429","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 work, we present a new sparse adaptive filtering algorithm following a variational Bayesian approach. First, sparsity is imposed by assigning Laplace priors to the filter parameters through a suitably defined hierarchical Bayesian model. Then, a variational Bayesian inference method is presented, which is appropriate for batch processing. In order to introduce adaptivity the Gauss-Seidel iterative scheme is properly embedded in our method. The proposed algorithm is fully automatic and is computationally efficient despite its Bayesian origin. Experimental results show that the algorithm converges to sparse solutions and exhibits superior estimation performance compared to related state-of-the-art schemes.
{"title":"Variational Bayesian sparse adaptive filtering using a Gauss-Seidel recursive approach","authors":"K. Themelis, A. Rontogiannis, K. Koutroumbas","doi":"10.5281/ZENODO.43475","DOIUrl":"https://doi.org/10.5281/ZENODO.43475","url":null,"abstract":"In this work, we present a new sparse adaptive filtering algorithm following a variational Bayesian approach. First, sparsity is imposed by assigning Laplace priors to the filter parameters through a suitably defined hierarchical Bayesian model. Then, a variational Bayesian inference method is presented, which is appropriate for batch processing. In order to introduce adaptivity the Gauss-Seidel iterative scheme is properly embedded in our method. The proposed algorithm is fully automatic and is computationally efficient despite its Bayesian origin. Experimental results show that the algorithm converges to sparse solutions and exhibits superior estimation performance compared to related state-of-the-art schemes.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"249 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":"126795031","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 article provides a methodology for designing a proposal distribution in the context of particle filtering for terrain navigation. The suggested method is based on the use of an importance distribution centered around an estimate of the maximum a posteriori (MAP). By assuming a Gaussian prior, we show that the computation of the MAP can be reduced to an optimization problem in a space of lower state dimension. Furthermore, we introduce a new method for choosing the covariance of the proposal. In this case, numerical experiments show that the method can improve upon classical sampling methods.
{"title":"Proposal distribution for particle filtering applied to terrain navigation","authors":"Achille Murangira, C. Musso","doi":"10.5281/ZENODO.43715","DOIUrl":"https://doi.org/10.5281/ZENODO.43715","url":null,"abstract":"This article provides a methodology for designing a proposal distribution in the context of particle filtering for terrain navigation. The suggested method is based on the use of an importance distribution centered around an estimate of the maximum a posteriori (MAP). By assuming a Gaussian prior, we show that the computation of the MAP can be reduced to an optimization problem in a space of lower state dimension. Furthermore, we introduce a new method for choosing the covariance of the proposal. In this case, numerical experiments show that the method can improve upon classical sampling methods.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"58 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":"126678180","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 proposes impact noise suppression with a new phase-based detection. Different from any other conventional algorithms which rely on magnitude-based or model-based detection, a new impact-noise detection utilizing phase linearity of the input noisy signal is developed based on an analysis of an impulse. Phase slopes of the input noisy signal is compared with an ideal phase slope obtained from the peak magnitude of the time-domain noisy signal. Phase unwrapping problem is alleviated by the use of a rotation vector of frequency domain components. Evaluation with PESQ for push-button clicks shows an improvement of 0.9 over a conventional noise suppressor for voice communication. Comparison of enhanced signal spectrogram with that of clean speech demonstrates superior enhanced signal quality. The proposed phase-based detection may be combined with an amplitude-based detection for more accurate results.
{"title":"Impact-noise suppression with phase-based detection","authors":"A. Sugiyama, Ryoji Miyahara, Kwangsoo Park","doi":"10.5281/ZENODO.43724","DOIUrl":"https://doi.org/10.5281/ZENODO.43724","url":null,"abstract":"This paper proposes impact noise suppression with a new phase-based detection. Different from any other conventional algorithms which rely on magnitude-based or model-based detection, a new impact-noise detection utilizing phase linearity of the input noisy signal is developed based on an analysis of an impulse. Phase slopes of the input noisy signal is compared with an ideal phase slope obtained from the peak magnitude of the time-domain noisy signal. Phase unwrapping problem is alleviated by the use of a rotation vector of frequency domain components. Evaluation with PESQ for push-button clicks shows an improvement of 0.9 over a conventional noise suppressor for voice communication. Comparison of enhanced signal spectrogram with that of clean speech demonstrates superior enhanced signal quality. The proposed phase-based detection may be combined with an amplitude-based detection for more accurate results.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"679 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":"116951884","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}
The problem of synchronization of Doppler-stretched GPS signals is addressed. A new synchronization technique for the course acquisition code is proposed, which does not assume that the so called narrow-band condition is satisfied. Thus, the Doppler effect is modeled as a frequency shift on the carrier and a stretching of the complex envelope. As a consequence, observation intervals significantly longer than those adopted in conventional techniques can be adopted leading to a higher immunity against disturbance signals. The proposed technique exploits the cyclostationarity of the received signal and the fact that transmitted and received signals are jointly spectrally correlated. Simulation results show the better performance of the proposed method with respect to the conventional technique that models the Doppler effect just as a frequency shift of the carrier.
{"title":"On synchronization of Doppler-stretched GPS signals","authors":"A. Napolitano, Ivana Perna","doi":"10.5281/ZENODO.43503","DOIUrl":"https://doi.org/10.5281/ZENODO.43503","url":null,"abstract":"The problem of synchronization of Doppler-stretched GPS signals is addressed. A new synchronization technique for the course acquisition code is proposed, which does not assume that the so called narrow-band condition is satisfied. Thus, the Doppler effect is modeled as a frequency shift on the carrier and a stretching of the complex envelope. As a consequence, observation intervals significantly longer than those adopted in conventional techniques can be adopted leading to a higher immunity against disturbance signals. The proposed technique exploits the cyclostationarity of the received signal and the fact that transmitted and received signals are jointly spectrally correlated. Simulation results show the better performance of the proposed method with respect to the conventional technique that models the Doppler effect just as a frequency shift of the carrier.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"41 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":"125247669","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}
A stereo blind source separation system for the iPhone is presented. A commercially available stereo microphone is installed on an iPhone and an input sound can be recorded by it or be transferred by PC through iTunes. The sound is then separated into two sources by a stereo-specific version of auxiliary-function-based independent vector analysis (Aux-IVA2) [1] in 1/5 the computational time of the input signal length. The quantitative evaluation for speech separation and other examples for source separation in real environments are presented.
{"title":"Blind source separation on iPhone in real environment","authors":"Nobutaka Ono","doi":"10.5281/ZENODO.43730","DOIUrl":"https://doi.org/10.5281/ZENODO.43730","url":null,"abstract":"A stereo blind source separation system for the iPhone is presented. A commercially available stereo microphone is installed on an iPhone and an input sound can be recorded by it or be transferred by PC through iTunes. The sound is then separated into two sources by a stereo-specific version of auxiliary-function-based independent vector analysis (Aux-IVA2) [1] in 1/5 the computational time of the input signal length. The quantitative evaluation for speech separation and other examples for source separation in real environments are presented.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"17 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":"114502141","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}
Mohamad El-Abed, A. Ninassi, C. Charrier, C. Rosenberger
The quality assessment of the acquired biometric raw data is very important as it deeply affects the performance of biometric systems and consequently their usability. Poor quality samples increase the enrolment failures, and decrease the system performance. In this paper, we present a new quality assessment metric of fingerprints. Its main originality lies in the use of a no-reference image quality metric. The proposed quality metric combines two types of parameters through a weighted sum optimized by a genetic algorithm: 1) image quality criterion and 2) pattern-based quality criteria (salient and patch-based features). BOZORTH3 matching system and the FVC2002 DB3 fingerprint database are used to clarify the benefits of the presented quality metric.
{"title":"Fingerprint quality assessment using a no-reference image quality metric","authors":"Mohamad El-Abed, A. Ninassi, C. Charrier, C. Rosenberger","doi":"10.5281/ZENODO.43594","DOIUrl":"https://doi.org/10.5281/ZENODO.43594","url":null,"abstract":"The quality assessment of the acquired biometric raw data is very important as it deeply affects the performance of biometric systems and consequently their usability. Poor quality samples increase the enrolment failures, and decrease the system performance. In this paper, we present a new quality assessment metric of fingerprints. Its main originality lies in the use of a no-reference image quality metric. The proposed quality metric combines two types of parameters through a weighted sum optimized by a genetic algorithm: 1) image quality criterion and 2) pattern-based quality criteria (salient and patch-based features). BOZORTH3 matching system and the FVC2002 DB3 fingerprint database are used to clarify the benefits of the presented quality metric.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"17 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":"123001213","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 proposes a tone-mapping method for High Dynamic Range (HDR) surveillance systems, which can improve the accuracy of SIFT object tracking in the systems. HDR surveillance systems can store minute details for the whole radiance range without over- or under-exposure. Although the minute details can be able to increase the accuracy of object detection and tracking, we have some issues of using them to existing SIFT. HDR contents should be tone-mapped before calculating SIFT feature since SIFT is designed for common Low Dynamic Range (LDR) contents. Most of conventional tone-mapping methods lose the minute details through the tone-mapping process, or degrade the accuracy of tracking since they also enhances the noise and invalid features. We propose the tone-mapping method for HDR surveillance system using SIFT features, which can enhance the minute details without invalid features.
{"title":"Tone-mapping for an HDR surveillance system using SIFT features","authors":"Takao Jinno, Shigeru Kuriyama, M. Okuda","doi":"10.5281/ZENODO.43733","DOIUrl":"https://doi.org/10.5281/ZENODO.43733","url":null,"abstract":"This paper proposes a tone-mapping method for High Dynamic Range (HDR) surveillance systems, which can improve the accuracy of SIFT object tracking in the systems. HDR surveillance systems can store minute details for the whole radiance range without over- or under-exposure. Although the minute details can be able to increase the accuracy of object detection and tracking, we have some issues of using them to existing SIFT. HDR contents should be tone-mapped before calculating SIFT feature since SIFT is designed for common Low Dynamic Range (LDR) contents. Most of conventional tone-mapping methods lose the minute details through the tone-mapping process, or degrade the accuracy of tracking since they also enhances the noise and invalid features. We propose the tone-mapping method for HDR surveillance system using SIFT features, which can enhance the minute details without invalid features.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"195 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":"122519617","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}
We present a novel method to exploit correlations of adjacent time-frequency (TF)-slots for a sparseness-based blind speech separation (BSS) system. Usually, these correlations are exploited by some heuristic smoothing techniques in the post-processing of the estimated soft TF masks. We propose a different approach: Based on our previous work with one-dimensional (1D)-hidden Markov models (HMMs) along the time axis we extend the modeling to two-dimensional (2D)-HMMs to exploit both temporal and spectral correlations in the speech signal. Based on the principles of turbo decoding we solved the complex inference of 2D-HMMs by a modified forward-backward algorithm which operates alternatingly along the time and the frequency axis. Extrinsic information is exchanged between these steps such that increasingly better soft time-frequency masks are obtained, leading to improved speech separation performance in highly reverberant recording conditions.
{"title":"Blind speech separation exploiting temporal and spectral correlations using 2D-HMMs","authors":"Dang Hai Tran Vu, Reinhold Häb-Umbach","doi":"10.5281/ZENODO.43429","DOIUrl":"https://doi.org/10.5281/ZENODO.43429","url":null,"abstract":"We present a novel method to exploit correlations of adjacent time-frequency (TF)-slots for a sparseness-based blind speech separation (BSS) system. Usually, these correlations are exploited by some heuristic smoothing techniques in the post-processing of the estimated soft TF masks. We propose a different approach: Based on our previous work with one-dimensional (1D)-hidden Markov models (HMMs) along the time axis we extend the modeling to two-dimensional (2D)-HMMs to exploit both temporal and spectral correlations in the speech signal. Based on the principles of turbo decoding we solved the complex inference of 2D-HMMs by a modified forward-backward algorithm which operates alternatingly along the time and the frequency axis. Extrinsic information is exchanged between these steps such that increasingly better soft time-frequency masks are obtained, leading to improved speech separation performance in highly reverberant recording conditions.","PeriodicalId":400766,"journal":{"name":"21st European Signal Processing Conference (EUSIPCO 2013)","volume":"106 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":"116107345","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：