By Topic

Digital Signal Processing, 2002. DSP 2002. 2002 14th International Conference on

Date 1-3 July 2002

Go

Filter Results

Displaying Results 1 - 25 of 210
  • 2002 14th International Conference on Digital Signal Processing Proceedings DSP

    Page(s): V - XLVIII
    Save to Project icon | Request Permissions | PDF file iconPDF (1035 KB)  
    Freely Available from IEEE
  • Author index

    Page(s): XLIX - LVI
    Save to Project icon | Request Permissions | PDF file iconPDF (311 KB)  
    Freely Available from IEEE
  • Robust and fast convergent blind multi-user detectors for DS-CDMA systems in nonstationary multipath environments

    Page(s): 627 - 630 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (294 KB) |  | HTML iconHTML  

    A set of adaptive receivers is proposed in this paper for direct-sequence code division multiple access (DS/CDMA) systems. Specifically we develop adaptive step-size blind algorithms which are based on constant modulus (CMA), constrained minimum output energy (CMOE) and multiuser constant modulus (MU-CMA) criteria. The applications of these algorithms to a nonstationary multipath environment were analyzed and simulation studies of the proposed algorithms were carried out. These studies show that the proposed adaptive step-size CMA, adaptive step-size CMOE and the adaptive step-size MU-CMA yield superior performance to most of the existing standard blind algorithms under severe nonstationary multipath fading channels. This is particularly the case when the number of interferers are nonstationary in the statistical sense. The robust nature of the proposed receivers is analysed also with respect to the initial step-size settings. In our simulations, both deterministic and Markovian time-varying interference statistics were considered. Specially, the Markovian time-varying case is of interest in mobile wireless systems where the number of users arrive according to a Poisson process, and hence, the interference statistics switch rapidly with time. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parametric architecture for implementing multimedia algorithms

    Page(s): 1261 - 1264 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (259 KB) |  | HTML iconHTML  

    Multimedia applications are characterized by high computational demands related to data transfer and storage operations. Multimedia algorithms in their majority consist of regular repetitive loop constructs. In this paper, a novel control architecture for implementing such loop intensive algorithms is described. The proposed control unit takes advantage of the regularity of computations in order to serve as high performance parametric controller of multimedia datapaths. The control unit cooperates with datapath modules and their corresponding controlling FSM. Algorithmic flow dependencies which determine the appropriate loop sequencing are mapped on a LUT. For another algorithm to execute, LUT context and FSM configurations only have to be reprogrammed. Thus, partial reconfiguration possibilities for implementing multimedia algorithms on programmable platforms can be exploited. For demonstration purposes, a matrix multiply algorithm implementation case is investigated. Compared to a software realization on ARM7 processor, significant performance improvements are reported. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A new CELP speech coder for wireless communication

    Page(s): 997 - 1000 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (320 KB) |  | HTML iconHTML  

    A new variable bit rate code excited linear predictive (CELP) coder of average bit rate around 4.2 kb/s is presented. The new coder is based on the federal standard FS 1016 CELP coder. Two main differences between this new coder and the FS 1016 coder are introduced. These two modifications will result in reducing the bit rate by about 0.53 kbits/s. A performance comparison between the proposed coder and other two coders is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improving Doppler ultrasound spectroscopy with multiband instantaneous energy separation

    Page(s): 611 - 614 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (348 KB) |  | HTML iconHTML  

    This paper deals with improving the time-frequency resolution of Doppler ultrasound spectroscopy, applied to blood flow analysis, by developing robust nonstationary spectrum estimation techniques based on Gabor (1946) filterbanks and multiband AM-FM demodulation that uses an instantaneous energy separation algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Feature extraction with wavelet transform for recognition of isolated handwritten Farsi/Arabic characters and numerals

    Page(s): 923 - 926 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (333 KB) |  | HTML iconHTML  

    A system is developed for recognition of handwritten Farsi/Arabic characters and numerals. The discrete wavelet transform is utilized to produce wavelet coefficients, which are used for classification. We used Haar wavelet for feature extraction in this system. The extracted features are used as training inputs to a feed forward neural network using the backpropagation learning rule. The learning and test patterns were gathered from various people with different educational backgrounds and different ages. We categorize 32 characters in Farsi language to 8 different classes in which characters of each class are very similar to each others. There are ten digits in Farsi/Arabic languages, but two of them are not used in postal codes in Iran, so we have 8 different extra classes for digits. This system yields the classification rates of 92.33% and 91.81% for these 8 classes of handwritten Farsi characters and numerals respectively. We used this system for recognizing the handwritten postal addresses which contain the names of cities and their postal codes. Our database contains 579 postal addresses in Iran. The system yields a recognition rate of 97.24% for these postal addresses. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • N dimensional Mojette transform. Application to multiple description

    Page(s): 1211 - 1214 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (331 KB)  

    The Mojette transform allows to exactly represent a discrete signal from a finite set of hyper-planes. This generic transform (many projection directions, choice of the number of projections, spline order) always ensures a very low complexity comparable to the FFT. In this paper, the Mojette transform in dimension n is presented and results issued from 2D and 3D cases are generalized. The central slice theorem, used with the Radon transform, is also derived in dimension n. Two schemes of applicatives examples enlight the interest for this generalization in the domain of multiple description. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On-road vehicle detection using Gabor filters and support vector machines

    Page(s): 1019 - 1022 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (482 KB) |  | HTML iconHTML  

    On-road vehicle detection is an important problem with application to driver assistance systems and autonomous, self-guided vehicles. The focus of this paper is on the problem of feature extraction and classification for rear-view vehicle detection. Specifically, we propose using Gabor filters for vehicle feature extraction and support vector machines (SVM) for vehicle detection. Gabor filters provide a mechanism for obtaining some degree of invariance to intensity due to global illumination, selectivity in scale, and selectivity in orientation. Basically, they are orientation and scale tunable edge and line detectors. Vehicles do contain strong edges and lines at different orientation and scales, thus, the statistics of these features (e.g., mean, standard deviation, and skewness) could be very powerful for vehicle detection. To provide robustness, these statistics are not extracted from the whole image but rather are collected from several subimages obtained by subdividing the original image into subwindows. These features are then used to train a SVM classifier. Extensive experimentation and comparisons using real data, different features (e.g., based on principal components analysis (PCA)), and different classifiers (e.g., neural networks (NN)) demonstrate the superiority of the proposed approach which has achieved an average accuracy of 94.81% on completely novel test images. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimizing the sweep cycle of time-varying comb filters for binaural dichotic presentation in sensorineural hearing impairment

    Page(s): 1145 - 1148 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (382 KB) |  | HTML iconHTML  

    Splitting of speech into two using time-varying comb filters has helped in reducing the effect of increased temporal and spectral masking simultaneously. The time varying comb filters contained pre-calculated sets of coefficients, which were selected in steps, such that a cyclic sweeping of magnitude responses occur. Presently the investigation has been carried out to find the best sweep cycle. The two sets of coefficients selected at any instant of time formed a pair of complementary comb filters with 18 auditory critical bands which provided spectral separation of components that are likely to get masked. Sweeping of the bands provides temporal separation and reduces the effect of temporal masking. As the number of magnitude responses swept increases, the sweeping becomes smoother. The sweep cycle duration determines the time for which the sensory cells of the basilar membrane remain stimulated. Perception evaluation tests were conducted with slowly sweeping sine wave and running speech from a male and a female speaker for sweep cycles of 10, 20, 40, 50, 60, 80, and 100 ms. The best results were obtained for a sweep cycle of 50 ms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Composite PNLMS & NLMS adaptation: a new method for network echo cancellation

    Page(s): 757 - 760 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (300 KB) |  | HTML iconHTML  

    This paper describes an improved version of the previously proposed fast converging algorithm, the proportionate normalized least mean squares (PNLMS), for network echo cancellers. We introduce a simple analysis of the PNLMS convergence behavior to show why after the fast initial convergence, it slows down. Also, the method has worked out to overcome this deficiency is presented. The improvement is shown by simulation results. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of expansion into matrix series to analysis of attractors of complex nonlinear dynamical systems

    Page(s): 959 - 962 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (273 KB) |  | HTML iconHTML  

    Decomposition methods of nonlinear operators describing the behavior of system in state space (phase space) are very important for analysis, identification and modeling of nonlinear dynamical systems (NDS), in particular NDS with self-organization (or complex NDS). The aim of this paper is derivation and classification of matrix series describing decomposition of vector functions from phase space variables and NDS operators into state space. This paper also develops some statements of matrix decomposition and main principles for analysis of attractors of complex NDS. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A prognostic-classification system based on a probabilistic NN for predicting urine bladder cancer recurrence

    Page(s): 1161 - 1164 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (509 KB) |  | HTML iconHTML  

    In this paper our purpose was to design a prognostic-classification system, based on a probabilistic neural network (PNN), for predicting urine bladder cancer recurrence. Ninety-two patients with bladder cancer were diagnosed and followed up. Images from each patient tissue sample were digitized and an adequate number of nuclei per case were segmented for the generation of morphological and textural nuclear features. Automatic urine bladder tumor characterization as a potential to recur or not was performed utilizing a PNN. An exhaustive search based on classifier performance indicated the best feature combination that produced the minimum classification error. The classification performance of the PNN was optimized employing a 4-dimensional feature vector that comprised one texture feature and three descriptors of nucleus size distribution. The classification accuracy for the group of cases with recurrence was 72.3% (35/47) and 71.1% (32/45) accuracy for the group of cases with no recurrence. The proposed prognostic-system could prove of value in rendering the diagnostic nuclear information a marker of disease recurrence. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On robust AR speech analysis based on quadratic classifier with heuristically decision threshold

    Page(s): 1077 - 1080 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (359 KB) |  | HTML iconHTML  

    The paper considers a robust recursive procedure for identifying a nonstationary AR speech model based on a quadratic classifier with a heuristic decision threshold. Two versions of the robust procedure with heuristic decision threshold, based on a frame-based quadratic classifier and a quadratic classifier with a sliding training data set, are evaluated and compared through analyzing natural speech signals with voiced and mixed excitation segments. The results obtained show that the considered robust procedure with the quadratic classifier with sliding training data set and heuristic decision threshold achieves more accurate AR speech parameter estimation, provides improved tracking performance, and achieves better discrimination capabilities for possible application in some vowel recognition systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of nonlinear signals based on estimating minimal attractor embedding dimension

    Page(s): 1001 - 1004 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (291 KB)  

    The theoretical ground of a local-topological method for defining minimal attractor embedding dimension on the basis of matrix decomposition for different types of dynamical systems is proposed. The computer confirmation of the theoretical results is presented. The investigation of digital electrocardiogram signals using local-topological analysis of chaotic attractor trajectories is carried out. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptable neural networks for modeling recursive non-linear systems

    Page(s): 1191 - 1194 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (311 KB) |  | HTML iconHTML  

    An efficient algorithm for recursive estimation of a non-linear autoregression (NAR) model is proposed. In particular, the model parameters are dynamically adapted through time so that (a) the model response, after the parameter updating, satisfies the current conditions and (b) a minimal modification of the model parameters is accomplished. The first condition is expressed by applying a first-order Taylor series to the non-linear function, which models the NAR system. The second condition implies the solution to be as much as close to the previous model state. The proposed recursive scheme is evaluated for traffic prediction of real-life MPEG coded video sources. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Space-Doppler parameter estimation for space-time-Doppler reception in time-varying multipath DS-CDMA systems

    Page(s): 1309 - 1312 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (346 KB) |  | HTML iconHTML  

    We propose a joint space-Doppler parameter estimation algorithm for fast-fading multipath DS-CDMA communication systems. The Doppler shifts induced by the relative motion between the transmitter and the receiver, together with the associated direction-of-arrival (DOA), are jointly estimated by a semi-blind subspace-type method. The Doppler effect is successfully eliminated by a space-time-Doppler receiver after the directional and Doppler parameters are estimated. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A scheduling scheme for multiplexing extra streaming data into digital TV programs

    Page(s): 579 - 582 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (344 KB) |  | HTML iconHTML  

    In digital TV systems, variable bitrate encoding is usually used to improve the bandwidth usage efficiency. Since the transmission channel has a fixed bandwidth, this leaves some portions of the bandwidth unoccupied. This free space can be used to transmit extra data to enhance the TV content, which can be either discrete, like text, or streaming, like video and audio. We address the problem of adding time-sensitive streaming data to a TV program. The crucial part of this problem is a scheduling algorithm that guarantees the on-time delivery of the incidental data to the decoder. We present a sophisticated time-sensitive scheduling algorithm for off-line multiplexing of TV programs and incidental streaming data. The two important features of our algorithm are: 1) it minimizes the presentation delay for incidental and main streams; 2) it minimizes the required decoder buffer size for incidental data. Comparing the experimental results of our algorithm with existing scheduling methods shows that our algorithm significantly reduces the presentation delay and the decoder buffer size for the incidental streams. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Scale hyperparameter estimation for GGMRF prior models with application to SPECT images

    Page(s): 521 - 524 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (308 KB) |  | HTML iconHTML  

    In this work we develop a Bayesian reconstruction method for SPECT (single photon emission computed tomography) images, using as prior GGMRF (generalized Gaussian Markov random fields) distributions and estimating the scale hyperparameter following the evidence analysis. Preconditioning methods are used to estimate this hyperparameter and the approximations used are compared on synthetic images. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • H bounded optimal updating - down-dating algorithm

    Page(s): 1293 - 1296 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (341 KB) |  | HTML iconHTML  

    The LMS algorithm, which is widely used in the adaptive filtering community, has been proved to be H optimal. We (Kothari et al. (2002)) have analyzed the other performance measures in the H setting which are of direct relevance to adaptive filtering and system identification. In that paper we considered the system identification and estimation employing exponential window problems. This problems are basically of rank I updating class, where we have to update the estimation as the new information comes into picture, while reducing the effect of the past data with a predefined factor. Due to this the effect of past data is not removed completely. The H performance measure in the situation of removing the past data effect completely and optimal H filter in this situation was still an open problem. In this paper we examine the performance measure in the H setting employing a sliding window. We present explicit algorithms and the achievable bound in this case. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Evaluating wavelet compression in microcalcification detection in mammography

    Page(s): 535 - 538 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (474 KB) |  | HTML iconHTML  

    The aim of this study is to evaluate the performance of a wavelet-based image compression algorithm in mammography. As microcalcifications detection is selected as diagnostic task, the algorithm is evaluated with respect to its visually lossless threshold at low compression ratios and not its supra-threshold performance. The threshold is determined by means of observer performance using a set of digitized mammograms from the Digital Database for Screening Mammography and by means of an image quality study, related to transfer characteristics of the compression algorithm, using computer-generated test images. The image quality parameters used are input/output response, noise, high contrast response and low contrast-detail response. The computer-generated test images mimic mammographic image and microcalcification characteristics. Receiver operating characteristics analysis of pooled data for microcalcification detection in mammograms indicated a threshold at compression ratio 40:1. Image quality parameters assessment, with respect to low contrast-detail patterns, indicated a threshold at compression ratio 35:1. The two approaches provided comparable thresholds, indicating the potential use of image quality parameters. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis and fast RLS algorithms of quadratic Volterra ADF

    Page(s): 749 - 752 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (316 KB) |  | HTML iconHTML  

    It is shown that adaptive training of quadratic Volterra filters is an ill-conditioned problem, or the error surfaces of the adaptive filters (ADF) are always extremely steep in one particular direction but relatively flat in the rest of the directions. This result is a generalization of a previous report on the special case of when the inputs are delayed values of a single time series of Gaussian distribution. A complete analysis of the correlation matrix of inputs as multiple time series are also obtained for the unrelated case. This paper then presents a fast RLS algorithm for Gaussian input signals costing only O(N2) multiplications where N is the number of linear terms in the filter input, the same order as the LMS algorithm, while the RLS algorithm for Volterra ADF costs O(N5) multiplications per sample. Simulations shown that this algorithm works well also in non-Gaussian input cases. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A robust speaker dependent algorithm for isolated word recognition

    Page(s): 993 - 996 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (318 KB) |  | HTML iconHTML  

    The paper deals with a simple speaker-dependent (SD) isolated word recognition (IWR) system based on template-based pattern matching. Different algorithms for storing and calculating the distortion between models and examples of words to be recognised are analysed. More specifically, the paper proposes a new algorithm that enhances performance with a slight increase in computational load and the amount of memory needed to store the models as compared with a traditional VQ-based approach. The results obtained in tests are given in terms of recognition rate, using the TIMIT-46 database with various type of background noise and different SNRs. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Single ideal-scale edge detection using noise within images

    Page(s): 935 - 938 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (579 KB) |  | HTML iconHTML  

    Researchers spend time and money developing techniques to solve some common problems associated with the rapidly developing field of image processing. These problems include poor edge detection in low contrast images, speed of recognition and high computational cost. Scale space analysis is an efficient solution to the edge detection of objects. However, this approach is time consuming and computationally expensive. These expenses can be marginally reduced if a single ideal (optimal) scale is found in scale space analysis and then edge detection is performed using only that single ideal scale. This paper reports on a new approach to detecting 3-dimensional objects in their 2-dimensional projections using noise within the images. The novel idea is based on selecting one ideal scale for the entire image at which edge detection can be applied. The selection of an ideal scale is based on the hypothesis that "an ideal edge detection scale depends on the noise within an image". This paper aims at presenting the above hypothesis mathematically and throughout some experiments made on simple 3-dimensional objects. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Heart period dynamics following an asphyxia experiment in rats

    Page(s): 1165 - 1168 vol.2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (316 KB) |  | HTML iconHTML  

    The present work proposes an optimal symbol dynamics transformation as an approach to analyze dynamic aspects of heart rate variability (HRV) in an asphyxia experiment in rats. The approximate entropy (ApEn) was used to quantify the regularity of short symbolic sequences derived from 5 min R-R intervals (RRI). The comparison of several symbol transformations applied in HRV analysis indicates our approach as optimal for characterizing the transitions between different phases in the recovery process following asphyxia in rats. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.