Pub Date : 1994-04-19DOI: 10.1109/ICASSP.1994.389991
Zhongkan Liu, Mingyong Zhou, H. Hama
In this paper the power spectral density of p-adic stationary stochastic process under the sense of Chrestenson transform (Ch-transform) and its maximum entropy estimator are studied. The relationship formula between the power spectral density and the entropy rate is first derived. The the normal equations of maximum Ch-entropy spectral estimator in closed expression are obtained. When the number of autocorrelation data is p/sup m/, where p/spl ges/2 and m/spl ges/1 are integers, the maximum Ch-entropy estimator can be directly expressed by the known finite autocorrelation data. These results are quite different from that of Fourier's. Numerical examples are provided to show the effectiveness of the maximum Ch-entropy estimator. General Hadmard ordering is introduced for the Kronecker formulation of the Ch-transform matrix. Such ordering can lead to a fast algorithm proposed in this paper which can reduce the computation complexity front O(p/sup 2m/) to O(mp/sup m/) when the number of autocorrelation data is p/sup m/ (m>1, p/spl ges/2).<>
{"title":"Maximum Ch-entropy estimation of p-adic stationary process and its fast algorithm","authors":"Zhongkan Liu, Mingyong Zhou, H. Hama","doi":"10.1109/ICASSP.1994.389991","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389991","url":null,"abstract":"In this paper the power spectral density of p-adic stationary stochastic process under the sense of Chrestenson transform (Ch-transform) and its maximum entropy estimator are studied. The relationship formula between the power spectral density and the entropy rate is first derived. The the normal equations of maximum Ch-entropy spectral estimator in closed expression are obtained. When the number of autocorrelation data is p/sup m/, where p/spl ges/2 and m/spl ges/1 are integers, the maximum Ch-entropy estimator can be directly expressed by the known finite autocorrelation data. These results are quite different from that of Fourier's. Numerical examples are provided to show the effectiveness of the maximum Ch-entropy estimator. General Hadmard ordering is introduced for the Kronecker formulation of the Ch-transform matrix. Such ordering can lead to a fast algorithm proposed in this paper which can reduce the computation complexity front O(p/sup 2m/) to O(mp/sup m/) when the number of autocorrelation data is p/sup m/ (m>1, p/spl ges/2).<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"15 12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127640623","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389440
R. Fioravanti, S. Fioravanti, D. Giusto
A novel predictive coding scheme for VQ is presented, called dynamic codebook reordering VQ (DCRVQ). Residual correlations between neighboring codevectors are exploited by a nonlinear prediction, that is a neural one. As a matter of fact, on the basis of the previously decoded codevectors, a multilayer neural network makes a prediction, and this result is used to reorganize the codebook in a dynamic way. This allows for efficient Huffman compression of codevector addresses after reordering.<>
{"title":"An efficient neural prediction for vector quantization","authors":"R. Fioravanti, S. Fioravanti, D. Giusto","doi":"10.1109/ICASSP.1994.389440","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389440","url":null,"abstract":"A novel predictive coding scheme for VQ is presented, called dynamic codebook reordering VQ (DCRVQ). Residual correlations between neighboring codevectors are exploited by a nonlinear prediction, that is a neural one. As a matter of fact, on the basis of the previously decoded codevectors, a multilayer neural network makes a prediction, and this result is used to reorganize the codebook in a dynamic way. This allows for efficient Huffman compression of codevector addresses after reordering.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127673601","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389762
I. Thng, A. Cantoni, Y. Leung
A method frequently employed to optimise the performance of a broadband antenna array processor is to find the set of processor weights which minimises its mean output power subject to linear constraints on the processor weights. These linear constraints determine the frequency response of the array processor in the look direction. To obtain robustness against mismatch between the array look direction and the actual signal direction, derivative constraints can be further imposed. However, derivative constraints corresponding to necessary and sufficient (NS) conditions for a maximally flat spatial power response in the look direction are in general quadratic. This limits their value for real-time processing. This paper presents several simpler derivative constraints. These constraints correspond only to sufficient conditions for maximal flatness but perform almost as well as the NS constraints and are better suited to real-time implementation.<>
{"title":"New sets of constraints for maximally flat optimum broadband antenna arrays","authors":"I. Thng, A. Cantoni, Y. Leung","doi":"10.1109/ICASSP.1994.389762","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389762","url":null,"abstract":"A method frequently employed to optimise the performance of a broadband antenna array processor is to find the set of processor weights which minimises its mean output power subject to linear constraints on the processor weights. These linear constraints determine the frequency response of the array processor in the look direction. To obtain robustness against mismatch between the array look direction and the actual signal direction, derivative constraints can be further imposed. However, derivative constraints corresponding to necessary and sufficient (NS) conditions for a maximally flat spatial power response in the look direction are in general quadratic. This limits their value for real-time processing. This paper presents several simpler derivative constraints. These constraints correspond only to sufficient conditions for maximal flatness but perform almost as well as the NS constraints and are better suited to real-time implementation.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126346098","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.390098
S. Qian, Dapang Chen
Using the orthogonal-like Gabor (1946) expansion, we decompose the Wigner-Ville (1932, 1948) distribution as a linear combination of localized and oscillated 2D Gaussian functions. Based on the degree of oscillation, we further group those 2D Gaussian functions into time-frequency functions P/sub d/(t,/spl omega/) and thereby obtain the so-called time-frequency distribution series (TFDS). The TFDS/sub D/ consists of up to Dth order P/sub d/(t,/spl omega/). Since the influence of P/sub d/(t,/spl omega/) on those useful properties is inversely proportional to d, by adjusting the order D of the TFDS one could easily balance the cross-term interference and the useful properties.<>
{"title":"Time-frequency distribution series","authors":"S. Qian, Dapang Chen","doi":"10.1109/ICASSP.1994.390098","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.390098","url":null,"abstract":"Using the orthogonal-like Gabor (1946) expansion, we decompose the Wigner-Ville (1932, 1948) distribution as a linear combination of localized and oscillated 2D Gaussian functions. Based on the degree of oscillation, we further group those 2D Gaussian functions into time-frequency functions P/sub d/(t,/spl omega/) and thereby obtain the so-called time-frequency distribution series (TFDS). The TFDS/sub D/ consists of up to Dth order P/sub d/(t,/spl omega/). Since the influence of P/sub d/(t,/spl omega/) on those useful properties is inversely proportional to d, by adjusting the order D of the TFDS one could easily balance the cross-term interference and the useful properties.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126382204","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389654
A. Kundu, George C. Chen, C. E. Persons
In ocean surveillance, a number of different types of transient signals are observed. These sonar signals are waveforms in one dimension (1-D), and often display an evolutionary pattern over the time scale. The hidden Markov model (HMM) is well-suited to classification of such 1-D signals. Following this intuition, the application of HMM to sonar transient classification is proposed and discussed in this paper. Toward this goal, three different feature vectors based on autoregressive (AR) model, Fourier power spectrum, and wavelet transforms are considered in our work. The neural net (NN) classifier has been successfully used for sonar transient classification. The same set of features as mentioned above is then used with an NN classifier. Some concrete experimental results using "DARPA standard data set I" with HMM and NN classification schemes are presented. Finally, a combined NN/HMM classifier is proposed, and its performance is evaluated with respect to individual classifiers.<>
{"title":"Transient sonar signal classification using hidden Markov model and neural net","authors":"A. Kundu, George C. Chen, C. E. Persons","doi":"10.1109/ICASSP.1994.389654","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389654","url":null,"abstract":"In ocean surveillance, a number of different types of transient signals are observed. These sonar signals are waveforms in one dimension (1-D), and often display an evolutionary pattern over the time scale. The hidden Markov model (HMM) is well-suited to classification of such 1-D signals. Following this intuition, the application of HMM to sonar transient classification is proposed and discussed in this paper. Toward this goal, three different feature vectors based on autoregressive (AR) model, Fourier power spectrum, and wavelet transforms are considered in our work. The neural net (NN) classifier has been successfully used for sonar transient classification. The same set of features as mentioned above is then used with an NN classifier. Some concrete experimental results using \"DARPA standard data set I\" with HMM and NN classification schemes are presented. Finally, a combined NN/HMM classifier is proposed, and its performance is evaluated with respect to individual classifiers.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126385100","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389634
B. Gordon, T. Meng
The paper describes a VLSI architecture designed to reconstruct a subband-encoded video stream. This architecture differs from previous designs in its low power operation. The chip operates at a maximum 20 MHz with a 1.5 V supply and can process up to 10 M color (YUV) pixels/sec while dissipating only 16 mW. The low power consumption is achieved through efficient algorithm-to-hardware mapping, a low-complexity subband filter, minimal memory accesses, a reduced supply voltage, and elimination of external memory support. A single chip will support color images up to 352 pixels wide, while multiple chip configurations can achieve any desired display resolution.<>
{"title":"A low power subband video decoder architecture","authors":"B. Gordon, T. Meng","doi":"10.1109/ICASSP.1994.389634","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389634","url":null,"abstract":"The paper describes a VLSI architecture designed to reconstruct a subband-encoded video stream. This architecture differs from previous designs in its low power operation. The chip operates at a maximum 20 MHz with a 1.5 V supply and can process up to 10 M color (YUV) pixels/sec while dissipating only 16 mW. The low power consumption is achieved through efficient algorithm-to-hardware mapping, a low-complexity subband filter, minimal memory accesses, a reduced supply voltage, and elimination of external memory support. A single chip will support color images up to 352 pixels wide, while multiple chip configurations can achieve any desired display resolution.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128137769","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389628
L. Guerra, M. Potkonjak, J. Rabaey
The exploration of concurrency has emerged as a dominant research problem in the VLSI DSP literature. A great variety of forms of concurrency exploration have been proposed, analyzed and used on a number of hardware platforms. The belief that the DSP domain is amenable to concurrency exploration has been gaining popularity; however, no systematic study has been conducted to confirm or dispel this claim. This paper presents a global view of the concurrency problem, presenting comprehensive statistics on concurrency properties in commonly used DSP programs. Particular emphasis is placed on the potential cost effectiveness of concurrency exploitation.<>
{"title":"Concurrency characteristics in DSP programs","authors":"L. Guerra, M. Potkonjak, J. Rabaey","doi":"10.1109/ICASSP.1994.389628","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389628","url":null,"abstract":"The exploration of concurrency has emerged as a dominant research problem in the VLSI DSP literature. A great variety of forms of concurrency exploration have been proposed, analyzed and used on a number of hardware platforms. The belief that the DSP domain is amenable to concurrency exploration has been gaining popularity; however, no systematic study has been conducted to confirm or dispel this claim. This paper presents a global view of the concurrency problem, presenting comprehensive statistics on concurrency properties in commonly used DSP programs. Particular emphasis is placed on the potential cost effectiveness of concurrency exploitation.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125752793","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389704
P. Mermelstein, Ping Zheng, M. Saikaly
We explore the benefits for CELP coding of speech at 8 kb/s of dividing the residual signal after pitch filtering into three band-passed components and using separate codebooks to represent each component. Minimization of the perceptually weighted error between the input signal and the reconstructed signal is divided into several band-limited minimization operations where the lowest frequency match dominates the quality of the result. For equal total numbers of bits allocated to code the residual in a 5 ms frame, spectral division of the coding operation results on the average in a better match than temporal division into subframes. These results permit the design of a high quality speech codec at 8 kb/s with modest delay and low complexity.<>
{"title":"Multi-band residual coding of CELP codecs at 8 kb/s","authors":"P. Mermelstein, Ping Zheng, M. Saikaly","doi":"10.1109/ICASSP.1994.389704","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389704","url":null,"abstract":"We explore the benefits for CELP coding of speech at 8 kb/s of dividing the residual signal after pitch filtering into three band-passed components and using separate codebooks to represent each component. Minimization of the perceptually weighted error between the input signal and the reconstructed signal is divided into several band-limited minimization operations where the lowest frequency match dominates the quality of the result. For equal total numbers of bits allocated to code the residual in a 5 ms frame, spectral division of the coding operation results on the average in a better match than temporal division into subframes. These results permit the design of a high quality speech codec at 8 kb/s with modest delay and low complexity.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125790627","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389607
Kamal Nourji, N. Demassieux
Digital signal processing algorithms often use inner product as basic computation. In this paper we propose a new design methodology for synthesizing an optimal VLSI architecture implementing a real-time Distributed Arithmetic-based inner product. Our design methodology considers the design space as bidimensional one. In the first dimension we consider all possible input data parallelisations: from bit-serial to bit-parallel. In the second dimension we consider all possible lookup-table partitioning. Using a new ROM generic model, expressions are developed for area and maximum input data bandwidth, which allows to have an explicit formulation of the area-bandwidth tradeoff. Finally, for a given set of application constraints (inner product size and data bandwidth), we exhibit the optimal architectural parameters that provide the smallest chip area.<>
{"title":"Optimal VLSI architecture for distributed arithmetic-based algorithms","authors":"Kamal Nourji, N. Demassieux","doi":"10.1109/ICASSP.1994.389607","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389607","url":null,"abstract":"Digital signal processing algorithms often use inner product as basic computation. In this paper we propose a new design methodology for synthesizing an optimal VLSI architecture implementing a real-time Distributed Arithmetic-based inner product. Our design methodology considers the design space as bidimensional one. In the first dimension we consider all possible input data parallelisations: from bit-serial to bit-parallel. In the second dimension we consider all possible lookup-table partitioning. Using a new ROM generic model, expressions are developed for area and maximum input data bandwidth, which allows to have an explicit formulation of the area-bandwidth tradeoff. Finally, for a given set of application constraints (inner product size and data bandwidth), we exhibit the optimal architectural parameters that provide the smallest chip area.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121840543","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 : 1994-04-19DOI: 10.1109/ICASSP.1994.389826
Hsien-Tsai Wu, J. Yang, Fwu-Kuen Chen
The eigenstructure based estimator designed to be used with the aid of the Gerschgorin's disk theorem is proposed for source number detection. By introducing the unitary transformation of the covariance matrix, the Gerschgorin radii of the eigenstructure are exploited to determine the number of sources while overcoming a lack of data samples, noise model and data independency information. Unlike conventional methods such as Akaike information criterion (AIC) and minimum descriptive length criterion (MDL), which are based on the cluster analysis of the eigenvalues used in conjunction with statistical formulations, the proposed method called the Gerschgorin disk estimator (GDE), provide more accurate detection of the source number in situations of both simulated and measured experimental data.<>
{"title":"Source number estimator using Gerschgorin disks","authors":"Hsien-Tsai Wu, J. Yang, Fwu-Kuen Chen","doi":"10.1109/ICASSP.1994.389826","DOIUrl":"https://doi.org/10.1109/ICASSP.1994.389826","url":null,"abstract":"The eigenstructure based estimator designed to be used with the aid of the Gerschgorin's disk theorem is proposed for source number detection. By introducing the unitary transformation of the covariance matrix, the Gerschgorin radii of the eigenstructure are exploited to determine the number of sources while overcoming a lack of data samples, noise model and data independency information. Unlike conventional methods such as Akaike information criterion (AIC) and minimum descriptive length criterion (MDL), which are based on the cluster analysis of the eigenvalues used in conjunction with statistical formulations, the proposed method called the Gerschgorin disk estimator (GDE), provide more accurate detection of the source number in situations of both simulated and measured experimental data.<<ETX>>","PeriodicalId":290798,"journal":{"name":"Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127993291","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：