Pub Date : 2000-06-05DOI: 10.1109/ICASSP.2000.859151
Sungjoo Ahn, Sunmee Kang, Hanseok Ko
This paper concerns effective speaker adaptation methods to solve the over-training problem in speaker verification, which frequently occurs when modeling a speaker with sparse training data. While various speaker adaptations have already been applied to speech recognition, these methods have not yet been formally considered in speaker verification. This paper proposes speaker adaptation methods using a combination of maximum a posteriori (MAP) and maximum likelihood linear regression (MLLR) adaptations, which are successfully used in speech recognition, and applies to speaker verification. Our aim is to remedy the small training data problem by investigating effective speaker adaptations for speaker modeling. Experimental results show that the speaker verification system using a weighted MAP and MLLR adaptation outperforms that of the conventional speaker models without adaptation by a factor of up to 5 times. From these results, we show that the speaker adaptation method achieves significantly better performance even when only small training data is available for speaker verification.
{"title":"Effective speaker adaptations for speaker verification","authors":"Sungjoo Ahn, Sunmee Kang, Hanseok Ko","doi":"10.1109/ICASSP.2000.859151","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.859151","url":null,"abstract":"This paper concerns effective speaker adaptation methods to solve the over-training problem in speaker verification, which frequently occurs when modeling a speaker with sparse training data. While various speaker adaptations have already been applied to speech recognition, these methods have not yet been formally considered in speaker verification. This paper proposes speaker adaptation methods using a combination of maximum a posteriori (MAP) and maximum likelihood linear regression (MLLR) adaptations, which are successfully used in speech recognition, and applies to speaker verification. Our aim is to remedy the small training data problem by investigating effective speaker adaptations for speaker modeling. Experimental results show that the speaker verification system using a weighted MAP and MLLR adaptation outperforms that of the conventional speaker models without adaptation by a factor of up to 5 times. From these results, we show that the speaker adaptation method achieves significantly better performance even when only small training data is available for speaker verification.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128792123","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.861068
G. Gelli, F. Verde
The problem of jointly equalizing a digital communication signal distorted by a linear time-invariant channel and rejecting co-channel or adjacent-channel digital interference is tackled. Owing to the presence of the interfering signal, the proposed optimum linear MMSE equalizer turns out to be periodically time-varying (LPTV). Moreover, new simple and effective blind channel identification procedures are presented, which can be applied as long as the desired and interfering signal exhibit different circularity and/or cyclo-stationarity properties. Simulation results confirm the effectiveness of the proposed techniques.
{"title":"Blind LPTV joint equalization and interference suppression","authors":"G. Gelli, F. Verde","doi":"10.1109/ICASSP.2000.861068","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.861068","url":null,"abstract":"The problem of jointly equalizing a digital communication signal distorted by a linear time-invariant channel and rejecting co-channel or adjacent-channel digital interference is tackled. Owing to the presence of the interfering signal, the proposed optimum linear MMSE equalizer turns out to be periodically time-varying (LPTV). Moreover, new simple and effective blind channel identification procedures are presented, which can be applied as long as the desired and interfering signal exhibit different circularity and/or cyclo-stationarity properties. Simulation results confirm the effectiveness of the proposed techniques.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"155 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128700980","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.859048
M. Doroslovački
Questions have previously been raised about the existence of an analytic signal with positive instantaneous frequency when the form of the analytic signal is prescribed. Here, it is shown that the complex function a(t)exp[j(/spl omega//sub 0/t+m(t))] is an analytic signal when m(t) is a real periodic function and a(t) is a band-limited real function with the maximum bandwidth depending on /spl omega//sub 0/ and the fundamental frequency of m(t). That implies as a special case m(t) which is simultaneously a periodic and piecewise polynomial. Positivity of the instantaneous frequency is simply obtained by requiring that the absolute value of the first derivative of m(t) is smaller than /spl omega//sub 0/.
{"title":"Nontrivial analytic signals with positive instantaneous frequency and band-limited amplitude","authors":"M. Doroslovački","doi":"10.1109/ICASSP.2000.859048","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.859048","url":null,"abstract":"Questions have previously been raised about the existence of an analytic signal with positive instantaneous frequency when the form of the analytic signal is prescribed. Here, it is shown that the complex function a(t)exp[j(/spl omega//sub 0/t+m(t))] is an analytic signal when m(t) is a real periodic function and a(t) is a band-limited real function with the maximum bandwidth depending on /spl omega//sub 0/ and the fundamental frequency of m(t). That implies as a special case m(t) which is simultaneously a periodic and piecewise polynomial. Positivity of the instantaneous frequency is simply obtained by requiring that the absolute value of the first derivative of m(t) is smaller than /spl omega//sub 0/.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"163 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129269795","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.860150
H. B. Yang, L. Jiao
Neural networks for solving a class of generalized eigenvalue problems of matrix pair are proposed, in which a universal function satisfying several conditions is introduced by replacing some ones. For its simplicity in structure and excellence in performance, it can be widely used in many areas including array signal processing, blind equalization and identification. Both the theoretical analysis and the experimental results show that the proposed network can gives the extreme eigenvalue and its corresponding eigenvector of the matrix pair (A, B) in real time.
{"title":"Neural network for solving generalized eigenvalues of matrix pair","authors":"H. B. Yang, L. Jiao","doi":"10.1109/ICASSP.2000.860150","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.860150","url":null,"abstract":"Neural networks for solving a class of generalized eigenvalue problems of matrix pair are proposed, in which a universal function satisfying several conditions is introduced by replacing some ones. For its simplicity in structure and excellence in performance, it can be widely used in many areas including array signal processing, blind equalization and identification. Both the theoretical analysis and the experimental results show that the proposed network can gives the extreme eigenvalue and its corresponding eigenvector of the matrix pair (A, B) in real time.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129305989","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.861148
S. Mudulodu, A. Paulraj
We propose a new blind space-time linear multiuser receiver for the CDMA downlink. The structure in the multiple access interference (MAI) for the downlink makes it possible for a receiver to suppress it without first obtaining the matched filter outputs for all the users (which comprise the sufficient statistics for optimum detection of the transmitted symbols); a 2D-RAKE receiver with appropriately chosen taps can suppress MAI reasonably well. Knowledge of all the users' codes is assumed and the taps of the 2D-RAKE receiver are estimated by making use of the subspace structure in the transmitted signal. Our approach does not require training symbols or channel estimation in order to estimate the receiver taps.
{"title":"A blind multiuser receiver for the CDMA downlink","authors":"S. Mudulodu, A. Paulraj","doi":"10.1109/ICASSP.2000.861148","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.861148","url":null,"abstract":"We propose a new blind space-time linear multiuser receiver for the CDMA downlink. The structure in the multiple access interference (MAI) for the downlink makes it possible for a receiver to suppress it without first obtaining the matched filter outputs for all the users (which comprise the sufficient statistics for optimum detection of the transmitted symbols); a 2D-RAKE receiver with appropriately chosen taps can suppress MAI reasonably well. Knowledge of all the users' codes is assumed and the taps of the 2D-RAKE receiver are estimated by making use of the subspace structure in the transmitted signal. Our approach does not require training symbols or channel estimation in order to estimate the receiver taps.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129733491","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.860198
D. Popescu, C. Rose
Programmable radios offer a new perspective on wireless communications since the modulation method is no longer fixed. Adaptive methods where user signatures and corresponding receiver filters are iteratively adapted can be used to improve performance. However, since codeword adjustments must be fed back to the transmitter, compact representation of codewords is extremely important. This issue is important for systems which employ interference avoidance since as opposed to current CDMA systems where uniform-amplitude codeword chips are used, interference avoidance employs real-valued "chips"-real-valued coefficients for a set of orthonormal basis functions of the signal space used by the transmitter and receiver. The paper represents a simple investigation of how codeword quantization affects the performance of interference avoidance algorithms. Results indicate that using 4-5 bits per chip for codeword representation is sufficient to maintain performance close to optimal values.
{"title":"Codeword quantization for interference avoidance","authors":"D. Popescu, C. Rose","doi":"10.1109/ICASSP.2000.860198","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.860198","url":null,"abstract":"Programmable radios offer a new perspective on wireless communications since the modulation method is no longer fixed. Adaptive methods where user signatures and corresponding receiver filters are iteratively adapted can be used to improve performance. However, since codeword adjustments must be fed back to the transmitter, compact representation of codewords is extremely important. This issue is important for systems which employ interference avoidance since as opposed to current CDMA systems where uniform-amplitude codeword chips are used, interference avoidance employs real-valued \"chips\"-real-valued coefficients for a set of orthonormal basis functions of the signal space used by the transmitter and receiver. The paper represents a simple investigation of how codeword quantization affects the performance of interference avoidance algorithms. Results indicate that using 4-5 bits per chip for codeword representation is sufficient to maintain performance close to optimal values.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126672391","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.860209
P. Vandenameele, L. Perre, B. Gyselinckx, M. Engels, M. Moonen, H. Man
Two major technical challenges in the design of future broadband wireless networks are the impairments of the propagation channel and the need for spectral efficiency. We previously proposed a combined OFDM/SDMA approach that mitigates the channel impairments by orthogonal frequency division multiplexing (OFDM) with cyclic prefix insertion and that achieves a high spectral efficiency by space division multiple access (SDMA). However, because of the multicarrier modulation, this approach requires high-backoff power amplifiers in the analog frontend. We present a SC-FD-SDMA basestation, which avoids these expensive amplifiers by using constant-envelope single-carrier (SC) modulation and still features the advantages of frequency-domain (FD) multipath mitigation and SDMA. We pay special attention to the initialization of such basestation and its fixed point requirements, since they are critical aspects of any realistic implementation. A case-study shows how SC-FD-SDMA enables a 100 Mbps wireless LAN with a bandwidth efficiency of 8 bps/Hz and an uncoded BER of 10/sup -3/ at 13.5 dB.
{"title":"A single-carrier frequency-domain SDMA basestation","authors":"P. Vandenameele, L. Perre, B. Gyselinckx, M. Engels, M. Moonen, H. Man","doi":"10.1109/ICASSP.2000.860209","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.860209","url":null,"abstract":"Two major technical challenges in the design of future broadband wireless networks are the impairments of the propagation channel and the need for spectral efficiency. We previously proposed a combined OFDM/SDMA approach that mitigates the channel impairments by orthogonal frequency division multiplexing (OFDM) with cyclic prefix insertion and that achieves a high spectral efficiency by space division multiple access (SDMA). However, because of the multicarrier modulation, this approach requires high-backoff power amplifiers in the analog frontend. We present a SC-FD-SDMA basestation, which avoids these expensive amplifiers by using constant-envelope single-carrier (SC) modulation and still features the advantages of frequency-domain (FD) multipath mitigation and SDMA. We pay special attention to the initialization of such basestation and its fixed point requirements, since they are critical aspects of any realistic implementation. A case-study shows how SC-FD-SDMA enables a 100 Mbps wireless LAN with a bandwidth efficiency of 8 bps/Hz and an uncoded BER of 10/sup -3/ at 13.5 dB.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126701805","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.860232
S. Godsill
We have previously shown how to perform inference about symmetric stable processes using Monte Carlo EM (MCEM) and Markov chain Monte Carlo (MCMC) techniques. Simulation based methods such as these are an excellent tool for inference with stable law distributions, since they do not require any direct evaluation of the stable density function, which is unavailable analytically in the general case. We review the existing methods for inference with MCMC and propose new methods based on the slice sampler, a very simple sampling algorithm which draws points from a uniform distribution over the area under the required density function. There is some evidence in the literature that the slice sampler has better convergence properties than the independence Metropolis samplers and rejection samplers previously proposed. We investigate this in the context of alpha-stable noise distributions.
{"title":"Inference in symmetric alpha-stable noise using MCMC and the slice sampler","authors":"S. Godsill","doi":"10.1109/ICASSP.2000.860232","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.860232","url":null,"abstract":"We have previously shown how to perform inference about symmetric stable processes using Monte Carlo EM (MCEM) and Markov chain Monte Carlo (MCMC) techniques. Simulation based methods such as these are an excellent tool for inference with stable law distributions, since they do not require any direct evaluation of the stable density function, which is unavailable analytically in the general case. We review the existing methods for inference with MCMC and propose new methods based on the slice sampler, a very simple sampling algorithm which draws points from a uniform distribution over the area under the required density function. There is some evidence in the literature that the slice sampler has better convergence properties than the independence Metropolis samplers and rejection samplers previously proposed. We investigate this in the context of alpha-stable noise distributions.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126903367","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.860170
A. Wahab, E. Tan, H. Abut
This paper focuses on a class of speech enhancement systems using the amplitude spectral estimates of noisy speech and noise to drive a Wiener filter to suppress simultaneously the ensemble of degradations picked up by microphones. A simple stereo microphone set, with left and right channels can be used to provide enough separation that unwanted signals can be reduced significantly to yield an acceptable quality speech signal for hands-free telephony applications in a vehicular environment. A generalized transform approach was introduced and experimental results show great potential when using the DCT as alternatives to the traditional Fourier transform approach to derive the amplitude spectral estimates of the corrupted speech signals and the noise process.
{"title":"Robust speech enhancement using amplitude spectral estimator","authors":"A. Wahab, E. Tan, H. Abut","doi":"10.1109/ICASSP.2000.860170","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.860170","url":null,"abstract":"This paper focuses on a class of speech enhancement systems using the amplitude spectral estimates of noisy speech and noise to drive a Wiener filter to suppress simultaneously the ensemble of degradations picked up by microphones. A simple stereo microphone set, with left and right channels can be used to provide enough separation that unwanted signals can be reduced significantly to yield an acceptable quality speech signal for hands-free telephony applications in a vehicular environment. A generalized transform approach was introduced and experimental results show great potential when using the DCT as alternatives to the traditional Fourier transform approach to derive the amplitude spectral estimates of the corrupted speech signals and the noise process.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129207191","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 : 2000-06-05DOI: 10.1109/ICASSP.2000.859260
Yuns Oh, Hong Jeong
We present a center-referenced basis for discrete representation of stereo correspondence that includes new occlusion nodes. This basis improves the inclusion of constraints and the parallelism of the final algorithm. Disparity estimation is formulated in a MAP context and natural constraints are incorporated, resulting in an optimal path problem in a sparsely connected trellis. Like other dynamic programming methods, the computational complexity is low at O(MN/sup 2/) for M/spl times/N pixel images. However, this method is better suited to parallel solution, scaling up to O(MN) processors. Experimental results confirm the performance of this method.
{"title":"Trellis-based parallel stereo matching","authors":"Yuns Oh, Hong Jeong","doi":"10.1109/ICASSP.2000.859260","DOIUrl":"https://doi.org/10.1109/ICASSP.2000.859260","url":null,"abstract":"We present a center-referenced basis for discrete representation of stereo correspondence that includes new occlusion nodes. This basis improves the inclusion of constraints and the parallelism of the final algorithm. Disparity estimation is formulated in a MAP context and natural constraints are incorporated, resulting in an optimal path problem in a sparsely connected trellis. Like other dynamic programming methods, the computational complexity is low at O(MN/sup 2/) for M/spl times/N pixel images. However, this method is better suited to parallel solution, scaling up to O(MN) processors. Experimental results confirm the performance of this method.","PeriodicalId":164817,"journal":{"name":"2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2000-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129290748","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：