Pub Date : 1991-05-09DOI: 10.1109/PACRIM.1991.160747
A. Richardson, T. Boswell, T.D. Blake
The Motorola DSP56000 and the Texas Instruments TMS320C25 digital signal processors have been used to implement real-time finite-impulse response (FIR) filters. Calculations were made to predict the number of instruction cycles needed to implement the FIR kernels of the assembly language programs, and the complete programs including input/output routines. Calculations based on the manufacturers' published data on instruction execution speeds showed that the FIR kernel of the DSP56000 assembly language program for an n-tap FIR would execute in (n+3) instruction cycles, whereas that of the TMS320C25 would execute in (n+7) instruction cycles. The calculations also predicted that, at a sampling rate of 48 kHz, the DSP56000 and the TMS320C25 would be able to implement up to 194-tap and 192-tap FIR filters, respectively. At 40 kHz sampling rate, a maximum of 234-tap FIR filter was calculated for the TMS320C25, and an experimental value of 227 was realized.<>
{"title":"Comparisons in performance between the Motorola DSP56000 and the Texas Instruments TMS320C25 for audio band, real-time digital signal processing","authors":"A. Richardson, T. Boswell, T.D. Blake","doi":"10.1109/PACRIM.1991.160747","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160747","url":null,"abstract":"The Motorola DSP56000 and the Texas Instruments TMS320C25 digital signal processors have been used to implement real-time finite-impulse response (FIR) filters. Calculations were made to predict the number of instruction cycles needed to implement the FIR kernels of the assembly language programs, and the complete programs including input/output routines. Calculations based on the manufacturers' published data on instruction execution speeds showed that the FIR kernel of the DSP56000 assembly language program for an n-tap FIR would execute in (n+3) instruction cycles, whereas that of the TMS320C25 would execute in (n+7) instruction cycles. The calculations also predicted that, at a sampling rate of 48 kHz, the DSP56000 and the TMS320C25 would be able to implement up to 194-tap and 192-tap FIR filters, respectively. At 40 kHz sampling rate, a maximum of 234-tap FIR filter was calculated for the TMS320C25, and an experimental value of 227 was realized.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115512669","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160702
L. Cossette, E. Velez, V. Cuperman
A discrete hidden Markov model (HMM) system recognizer using word-specific vector quantization is described. The word-specific VQ approach is suggested as an alternative to universal codebook vector quantization. The set of word-specific VQ index sequences is connected to each of the word-specific HMM models. For speaker-independent isolated digit recognition with a studio recorded database, a performance of 99.5% was obtained for the word-specific codebook VQ-HMM recognizer, which is an improvement of 2% when compared to a universal codebook VQ-HMM recognizer tested on the same speech database.<>
{"title":"Speaker-independent isolated-digit recognition based on hidden Markov models and multiple vocabulary specific vector quantization","authors":"L. Cossette, E. Velez, V. Cuperman","doi":"10.1109/PACRIM.1991.160702","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160702","url":null,"abstract":"A discrete hidden Markov model (HMM) system recognizer using word-specific vector quantization is described. The word-specific VQ approach is suggested as an alternative to universal codebook vector quantization. The set of word-specific VQ index sequences is connected to each of the word-specific HMM models. For speaker-independent isolated digit recognition with a studio recorded database, a performance of 99.5% was obtained for the word-specific codebook VQ-HMM recognizer, which is an improvement of 2% when compared to a universal codebook VQ-HMM recognizer tested on the same speech database.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114672085","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160810
S. Sayad, J. Sayad, M. Sayad
NN-VEIN is a novel approach to the structure of neural networks. In this structure the input units are located in a network of connections. Traditional NN and NN-VEIN are compared. In this structure, input units are located in weighted connections of which the degree may be variable from 0 to N. This degree varies with respect to the quantity of input data. This network is being used to design intelligent systems in medical diagnosis. The input layer in NN-VEIN in comparison with NN is completely different and is a really intelligent process. Clinical diagnosis expert systems created by the use of this network do not need an input pattern to be entered at once, but like an expert physician, any question is based on previous questions.<>
{"title":"Neural network with variable excitability of input units (NN-VEIN)","authors":"S. Sayad, J. Sayad, M. Sayad","doi":"10.1109/PACRIM.1991.160810","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160810","url":null,"abstract":"NN-VEIN is a novel approach to the structure of neural networks. In this structure the input units are located in a network of connections. Traditional NN and NN-VEIN are compared. In this structure, input units are located in weighted connections of which the degree may be variable from 0 to N. This degree varies with respect to the quantity of input data. This network is being used to design intelligent systems in medical diagnosis. The input layer in NN-VEIN in comparison with NN is completely different and is a really intelligent process. Clinical diagnosis expert systems created by the use of this network do not need an input pattern to be entered at once, but like an expert physician, any question is based on previous questions.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114693721","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160753
E.A. Quincy
Two objective measures, computed from received digital images, that can be used to assess video quality are proposed. These measures, based on Laplacian and Sobel gradient operators, are developed and applied to video teleconferencing (VTC) sequences of live desktop scenes and conference table scenes with motion present. The measured results are then compared over a range of transmission bit rates. The sensitivity and reliability of the measures are enhanced by determining the optimum gray level thresholds for computing the measures. Both measures are shown to monotonically increase with transmission rate and corresponding illustrative subjective evaluation of the images. These measures have been proposed for possible inclusion into a draft National VTC Standard under development in the ANSI T1Q1.5 teleconferencing standards group.<>
{"title":"Gradient measures of compressed video teleconferencing quality","authors":"E.A. Quincy","doi":"10.1109/PACRIM.1991.160753","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160753","url":null,"abstract":"Two objective measures, computed from received digital images, that can be used to assess video quality are proposed. These measures, based on Laplacian and Sobel gradient operators, are developed and applied to video teleconferencing (VTC) sequences of live desktop scenes and conference table scenes with motion present. The measured results are then compared over a range of transmission bit rates. The sensitivity and reliability of the measures are enhanced by determining the optimum gray level thresholds for computing the measures. Both measures are shown to monotonically increase with transmission rate and corresponding illustrative subjective evaluation of the images. These measures have been proposed for possible inclusion into a draft National VTC Standard under development in the ANSI T1Q1.5 teleconferencing standards group.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114933455","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160751
K. G. Kelley
The technical areas currently limiting video teleconferencing's (VTC's) effectiveness for meeting essential US Department of Defence (DoD) needs are identified, and the DoD response being taken to each is described. Special emphasis is given to a discussion of performance issues from a user's perspective. Suggestions are offered for resolving these issues within the framework of current commercial standards development activity.<>
{"title":"The need for video standards in DoD","authors":"K. G. Kelley","doi":"10.1109/PACRIM.1991.160751","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160751","url":null,"abstract":"The technical areas currently limiting video teleconferencing's (VTC's) effectiveness for meeting essential US Department of Defence (DoD) needs are identified, and the DoD response being taken to each is described. Special emphasis is given to a discussion of performance issues from a user's perspective. Suggestions are offered for resolving these issues within the framework of current commercial standards development activity.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117205319","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160743
Z. Li, R. Krishnan, R. Marks
An investigation is conducted into residue number system (RNS) to decimal number system conversion algorithms and their implementations. For general Q-tuple RNS decoding, the authors propose a modularized decoding algorithm in which each module decodes only a 2-tuple RNS code, and they discuss its computational complexity at the algorithm level. The second algorithm is for decoding a 2-tuple RNS code and can be called a partial table-lookup method. By memorizing only partial decimal numbers in the dynamic range of any RNS, this method can generate the correct decimal number with at most three additions. Hardware implementation methods are discussed, and a comparison of the algorithms with conventional decoding algorithms is made in terms of the arithmetic operation complexity, dynamic range requirements, and hardware implementation complexity.<>
{"title":"Modulized RNS-decimal number conversion algorithm and its implementations","authors":"Z. Li, R. Krishnan, R. Marks","doi":"10.1109/PACRIM.1991.160743","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160743","url":null,"abstract":"An investigation is conducted into residue number system (RNS) to decimal number system conversion algorithms and their implementations. For general Q-tuple RNS decoding, the authors propose a modularized decoding algorithm in which each module decodes only a 2-tuple RNS code, and they discuss its computational complexity at the algorithm level. The second algorithm is for decoding a 2-tuple RNS code and can be called a partial table-lookup method. By memorizing only partial decimal numbers in the dynamic range of any RNS, this method can generate the correct decimal number with at most three additions. Hardware implementation methods are discussed, and a comparison of the algorithms with conventional decoding algorithms is made in terms of the arithmetic operation complexity, dynamic range requirements, and hardware implementation complexity.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129875760","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160707
H. Yao, R. Podhorodeski, Z. Dong
A new analytical model and a range finder simulated on the basis of this model is presented. The optical paths of a range finder, both outside and inside of the device, are modeled using geometric transformations. The simulated range data are created using the 3-D geometric models of the scene, the new analytical model for range-finding devices, and the modified Z-buffer algorithm. The intent is to develop a flexible research tool for testing shape recognition algorithms of a computer vision system, and for providing guidelines for range-finding device design.<>
{"title":"Modelling and simulation of a range-finding device","authors":"H. Yao, R. Podhorodeski, Z. Dong","doi":"10.1109/PACRIM.1991.160707","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160707","url":null,"abstract":"A new analytical model and a range finder simulated on the basis of this model is presented. The optical paths of a range finder, both outside and inside of the device, are modeled using geometric transformations. The simulated range data are created using the 3-D geometric models of the scene, the new analytical model for range-finding devices, and the modified Z-buffer algorithm. The intent is to develop a flexible research tool for testing shape recognition algorithms of a computer vision system, and for providing guidelines for range-finding device design.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129377969","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160737
H. Haniuda, Y. Abiru, N. Miyazaki
PHI is an experimental distributed deductive database system. It is implemented on personal sequential inference machines (PSIs). A deductive database consists of intensional and extensional databases. PHI is composed of two layers which handle intensional and extensional databases, respectively. The query language of PHI is Datalog and its query processing is performed in a bottom-up manner. Query optimization of PHI is the combination of query transformation and techniques developed for relational databases. A discussion is presented of the design, implementation, and evaluation of the deductive database part of PHI.<>
{"title":"PHI: a deductive database system","authors":"H. Haniuda, Y. Abiru, N. Miyazaki","doi":"10.1109/PACRIM.1991.160737","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160737","url":null,"abstract":"PHI is an experimental distributed deductive database system. It is implemented on personal sequential inference machines (PSIs). A deductive database consists of intensional and extensional databases. PHI is composed of two layers which handle intensional and extensional databases, respectively. The query language of PHI is Datalog and its query processing is performed in a bottom-up manner. Query optimization of PHI is the combination of query transformation and techniques developed for relational databases. A discussion is presented of the design, implementation, and evaluation of the deductive database part of PHI.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127128461","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160745
R. Sivakumar, N.J. Dimopulos, K. Li
A VLSI design of a modulo-extractor based on the principles of residue arithmetic is discussed. A method for computing (X)/sub m/ for specific values of m is analyzed, and the area-time complexity has been implemented in 3 mu m CMOS3DLM technology. Simulation results have yielded a propagation delay of less than 100 ns.<>
讨论了基于残数算法原理的模提取器的VLSI设计。分析了特定m值(X)/sub m/的计算方法，并在3 μ m CMOS3DLM技术上实现了面积-时间复杂度。仿真结果显示，传输延迟小于100纳秒。
{"title":"VLSI design of a modulo-extractor","authors":"R. Sivakumar, N.J. Dimopulos, K. Li","doi":"10.1109/PACRIM.1991.160745","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160745","url":null,"abstract":"A VLSI design of a modulo-extractor based on the principles of residue arithmetic is discussed. A method for computing (X)/sub m/ for specific values of m is analyzed, and the area-time complexity has been implemented in 3 mu m CMOS3DLM technology. Simulation results have yielded a propagation delay of less than 100 ns.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130255335","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 : 1991-05-09DOI: 10.1109/PACRIM.1991.160676
J. Schroeder, R. Kubichek
The behavior of the CD (cepstral distance) is studied under serveral types of controlled distortion. In addition, CD values are computed from both L/sub 1/ and L/sub 2/ normed LPC (linear predictive coding) models; these results are compared and discussed. The analysis indicates that the L/sub 1/-based CD is more robust than more typical L/sub 2/-based solutions, especially when impulsive-type noise is present. The CD does not appear to be effective for echo-type distortions. However, it provided monotonic behavior for other impairments such as added noise, clipping and low-pass filtering.<>
{"title":"L/sub 1/ and L/sub 2/ normed cepstral distance controlled distortion performance (speech quality)","authors":"J. Schroeder, R. Kubichek","doi":"10.1109/PACRIM.1991.160676","DOIUrl":"https://doi.org/10.1109/PACRIM.1991.160676","url":null,"abstract":"The behavior of the CD (cepstral distance) is studied under serveral types of controlled distortion. In addition, CD values are computed from both L/sub 1/ and L/sub 2/ normed LPC (linear predictive coding) models; these results are compared and discussed. The analysis indicates that the L/sub 1/-based CD is more robust than more typical L/sub 2/-based solutions, especially when impulsive-type noise is present. The CD does not appear to be effective for echo-type distortions. However, it provided monotonic behavior for other impairments such as added noise, clipping and low-pass filtering.<<ETX>>","PeriodicalId":289986,"journal":{"name":"[1991] IEEE Pacific Rim Conference on Communications, Computers and Signal Processing Conference Proceedings","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128840799","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：