By Topic

Time-Frequency and Time-Scale Analysis, 1998. Proceedings of the IEEE-SP International Symposium on

Date 9-9 Oct. 1998

Filter Results

Displaying Results 1 - 25 of 170
  • Proceedings of the IEEE-SP International Symposium on Time-Frequency and Time-Scale Analysis (Cat. No.98TH8380)

    Publication Year: 1998
    Save to Project icon | Request Permissions | PDF file iconPDF (809 KB)  
    Freely Available from IEEE
  • Author index

    Publication Year: 1998 , Page(s): xii - xiv
    Save to Project icon | Request Permissions | PDF file iconPDF (160 KB)  
    Freely Available from IEEE
  • A wideband time-frequency Weyl symbol and its generalization

    Publication Year: 1998 , Page(s): 29 - 32
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (412 KB)  

    We extend the work of Shenoy and Parks (1994) on the wideband Weyl correspondence. We define a wideband Weyl symbol (P0WS) in the time-frequency plane based on the Bertrand (1988) P0-distribution, and we study its properties, examples and possible applications. Using warping relations, we generalize the P0WS and the wideband spreading function (WSF) to analyze systems producing dispersive time shifts. We provide properties and special cases (e.g. power and exponential) to demonstrate the importance of our generalization. The new generalized WSF provides a new interpretation of a system output as a weighted superposition of dispersive time-shifted versions of the signal. We provide application examples in analysis and detection to demonstrate the advantages of our new results for linear systems with group delay characteristics matched to the specific warping used View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • System level design of a pattern recognition system based on the Gabor wavelets

    Publication Year: 1998 , Page(s): 237 - 240
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (440 KB)  

    In this paper we will address the system level design of pattern recognition architectures. The system targeted uses Gabor wavelets for the multi-frequency and multi-scale analysis of gray scale images. The macro-blocks or macro-functions architectures are developed in sight of the integration of a system-on-a-chip. We have faced two basic design problems. First, to which extent the algorithms can be simplified to fit a hardware implementation? Second, how the functionality will be partitioned to fit an optimal hardware/software system? To carry out these operations we have relied on a design methodology taking into account the realization constraints early in the design stages. We used a fast cycle precise simulator to attempt and evaluate many hardware/software alternatives View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A nonuniform, shift-invariant, and optimal algorithm for Malvar's wavelet decomposition

    Publication Year: 1998 , Page(s): 45 - 48
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (268 KB)  

    In this article, we propose a new decomposition scheme for Malvar's wavelet representation. Our algorithm is nonuniform, shift-invariant and minimal for an information cost function, contrary to the shift-invariant algorithm of Cohen. We propose some restrictions to our algorithm in order to reduce the complexity and permitting us to provide some partitions of the signal in agreement with its structure. This new local trigonometric transform, more adapted than Malvar's decomposition, allows the analysis of the signal and permits one to obtain a satisfying time-frequency representation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Denoising of non-stationary signals using optimized Karhunen-Loeve expansion

    Publication Year: 1998 , Page(s): 621 - 624
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (256 KB)  

    We show that denoising a non-stationary signal is possible by means of a Karhunen-Loeve (KL) expansion optimized by an entropy criterion. This criterion is used to segment the noisy signal and to choose the most parsimonious KL representation possible for each segment. The entropy of the KL coefficients for different window lengths determines the appropriate number and the lengths of the windows. To find the KL coefficients in each segment, a time-varying autocorrelation matrix is estimated using the evolutionary periodogram. Eigenvalues and eigenvectors needed in the expansion are computed from this matrix. The local eigenvectors are the basis for each segment. An estimate of the evolutionary spectrum of the signal is obtained from the KL expansion. Choosing the KL coefficients corresponding to the most significant eigenvalues and time-windowing are shown to constitute masking in the time-frequency plane. This masking permits the denoising of non-stationary signals corrupted by white noise View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The multidimensional multi-window nonrectangular discrete Gabor schemes

    Publication Year: 1998 , Page(s): 41 - 44
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (220 KB)  

    The multi-window Gabor (1946) type scheme is generalized to the multidimensional case with special emphasis on nonrectangular lattices. The formulations are presented by means of matrix operations, which incorporate periodicity and sampling matrix that sample the continuous signal over different grids, such as hexagonal grids. The proposed scheme is discussed in the context of oversampling. Algorithms are presented in the Zak transform domain View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quantized high resolution pursuit

    Publication Year: 1998 , Page(s): 189 - 192
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (236 KB)  

    Because of their properties, adaptive decomposition techniques have been hardly studied in order to apply them to different works, such as signal characterization, compact representation, feature extraction, noise subtraction, etc. High resolution pursuit is one of the techniques developed which principally provides a very good super-resolution without requiring expensive computation, compared with other super-resolution techniques such as basis pursuit. When an adaptive decomposition technique is employed for compact representation, it is necessary to consider the introduction of a quantization stage for better results. V. Goyal (see SPIE vol.3024, p.2-12) presented good result and has also studied quantization in adaptive decomposition techniques, specifically with the matching pursuit. We test quantization with the high resolution pursuit technique in order to analyze some of the results obtained View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Time-frequency analysis of harmonic oscillator motion

    Publication Year: 1998 , Page(s): 25 - 28
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (260 KB)  

    Single and coupled harmonic oscillators have been used to model many physical systems in all fields of science. We show here that an effective way to study their motion is in the time-frequency plane. We numerically integrate a number of cases with various input forces and show that the time-frequency plane displays the oscillations in a clear and easily interpretable manner View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The higher order statistics of energy operators with application to neurological signals

    Publication Year: 1998 , Page(s): 561 - 564
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (360 KB)  

    Statistics for detecting changes in signal energy are developed for generalized energy estimation algorithms. The Teager energy operator (TEO) is a method for quantifying signal energy, a product of both frequency as well as amplitude. Using second and third order autocorrelation-based tests for dependence, we examine time domain methods of energy detection of sinusoids. To quantify signal energy we exploit the whiteness of the output of the TEO. The C-statistics examine the level of second order whiteness in a time series. The newly developed H-statistics test confirms the presence of third order whiteness or independence. A pure noise exhibits both second and third order whiteness. A power analysis of these tests for energy detection are also shown to be sensitive to changes in both sinusoidal amplitude and frequency. The Cand H-statistics allow for quantification of distortion in the TEO output as well. Distortion in an energy operator results from poor cancellation of cross-terms or from second harmonic distortion as typified by a traditional square law device. Fluctuations in band-specific EEG (electroencephalogram) energy also are amenable to practical analysis using the TEO. An example of an EEG signal with a large harmonic content are spindle signals taken from animal experiments dealing with recovery from hypoxic-asphyxic injury View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Signal synthesis and positive time frequency distributions

    Publication Year: 1998 , Page(s): 17 - 20
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (304 KB)  

    A method for obtaining a generalized transfer function (GTF) of a linear time-varying system based on the evolutionary spectral theory is proposed. This GTF can be used to synthesize the signal, and its magnitude squared function results in the signal's positive time frequency distribution that satisfies the marginals (i.e., a Cohen-Posch (1995) TFD). The procedure allows any prior estimate of the GTF to be modified such that the resulting posterior GTF is closest in the least square sense to the prior and satisfies the above mentioned properties. Examples are presented to illustrate the performance of the method View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Denoising using wavelet packets and the kurtosis: application to transient detection

    Publication Year: 1998 , Page(s): 625 - 628
    Cited by:  Papers (8)  |  Patents (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (408 KB)  

    The problem addressed in this paper is the detection of an unknown transient signal corrupted by additive Gaussian noise. We have shown in a previous study that Malvar wavelets can be successfully used when the noise is white Gaussian. The criterion to choose the best basis is based on the Gaussianity of the wavelet coefficients: when two adjacent segments have Gaussian coefficients they are merged, otherwise they are kept separated. If the noise is colored, this criterion fails to give good results. For this case we use wavelet packets instead. The best basis is chosen in the same way: merging “Gaussian frequency bands”. To get a time dependent detection statistic, we perform a denoising: Gaussian wavelet coefficients are set to zero. After reconstruction of the denoised signal, a standard detection procedure is performed. The performances of this detection scheme are studied experimentally. Furthermore, an application of the method is described for a real case View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Morphology-based perfect reconstruction filter banks

    Publication Year: 1998 , Page(s): 353 - 356
    Cited by:  Papers (2)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (268 KB)  

    This paper discusses the construction of perfect reconstruction filter banks using morphological operators. Two concrete examples are given: (i) the morphological Haar wavelet, and (ii) a wavelet decomposition obtained from the lifting scheme which has the nice property that local maxima are preserved View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Unhierarchical coding of image wavelet coefficients without sending the sign

    Publication Year: 1998 , Page(s): 465 - 468
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (264 KB)  

    A new algorithm for embedded image coding is presented. The method used is based on wavelet decomposition. We obtain results only slightly worse than those obtained using the best published algorithms based on the wavelet transform and from these results we assign a relative importance to the relation between the coefficients of different frequency bands. The main characteristics of this new codification are embedded code and coding the transform matrix without sending the sign of each coefficient and without keeping in mind any relation between the coefficients. We apply a successive quantization to the wavelet coefficients and an index to those coefficients greater than the threshold is sent to the decoder View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dynamical spectrograms that can be perceived as visual gestures

    Publication Year: 1998 , Page(s): 537 - 540
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (512 KB)  

    A new system for speech visualisation, has been implemented to allow deaf and hearing-impaired people to understand verbal information over channels such as the ordinary public telephone system. Incorporating a computational model of the human ear, the system converts incoming sounds into a sequence of animated images, which show the temporal variations of the spectral pattern of the input sound in real-time and which are perceived like visual gestures. Preliminary results from forced-choice tests with 28 human subjects are reported, using a sequence of 2- to 4-word sets. To demonstrate the language independence of this approach, some of these were taken from 4 very different languages-English, Persian, French and Czech. The results show high levels of recognition after only 10 learning trials (typical mean scores of 50-85%, where zero represents chance expectation), and encourage further investigation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Enhanced PN code tracking and detection using wavelet packet denoising

    Publication Year: 1998 , Page(s): 365 - 368
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (360 KB)  

    In an effort to transmit information over communication channels that contain both Gaussian and non-Gaussian noise components, military and commercial applications frequently rely upon the use of direct sequence spread spectrum (DSSS) techniques. Whilst DSSS is an interference tolerant modulation scheme, situations may be encountered when the processing gain offered by the pseudonoise (PN) spreading code is inadequate. Enhancements to the interference immunity of a DSSS system can be accomplished through the use of signal processing techniques that suppress unwanted interference prior to despreading. To date, the majority of interference suppression techniques developed for DSSS applications, have concentrated on improving detector performance exclusively, and perfect synchronisation between transmitter and receiver is assumed. However, the process of PN code acquisition and tracking within a DSSS receiver is arguably one of the most critical functions within the receiver structure since, if it fails to operate correctly, the wanted signal will not be successfully recovered. An evaluation of the benefits of any interference suppression technique within a DSSS receiver would not be complete without considering at least one level within the synchronisation hierarchy. We examine the improvements in receiver performance, when the process of PN code tracking and detection employs a novel wavelet packet denoising (WPD) procedure for a variety of non-Gaussian channel scenarios View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The discrete wavelet transform as a tool for automatic phase pickers

    Publication Year: 1998 , Page(s): 201 - 204
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (292 KB)  

    Seismic data consist of traces, which contain information about a seismic event, but in some period of time may virtually be noise. A trace, which contains seismic information, is called a seismic signal. Seismic signals consist of several typically short energy bursts, called phases, exhibiting several patterns in terms of dominant frequency, amplitude and polarization. Amongst others, significant phases are the P-phase and the S-phase. We present a fast algorithm to detect the S-phase in a seismic signal. In this method we use a combination of traditional S-phase detection methods from seismology and the discrete wavelet transform. First results are presented to demonstrate our new approach View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • New pitch detection algorithm based on wavelet transform

    Publication Year: 1998 , Page(s): 165 - 168
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (332 KB)  

    A new pitch detection algorithm based on wavelet transform analysis is presented. This algorithm uses a family of modulated Gaussian wavelets adapted to the Bark scale to analyse speech signals decomposing the input signal into different bands. Then, a maxima detector and a new confirmation algorithm are used to extract pitch period information. Evaluation results and comparison tests with standard SIFT algorithm are presented View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A comparison between conventional and wavelet based amplitude compression schemes

    Publication Year: 1998 , Page(s): 161 - 164
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (344 KB)  

    Loudness recruitment is a symptom of sensorineural hearing loss affecting the inner ear when the threshold of hearing is raised. The increase of the hearing threshold is often nonuniform across the range of audible frequencies and loudness perceptions are distorted. One method used to compensate for this type of hearing disorder is amplitude compression followed by equalization. In the present study, we compared amplitude compression in the discrete domain via two methods: (1) conventional two-channel amplitude compression, as proposed by Villchur and (2) wavelet based compression scheme using three levels of decomposition/reconstruction with the DB-9 wavelet. Both of these algorithms were tested on normal hearing subjects with elevated thresholds simulated by masking noise. The subjects were tested using nonsense sentences and the merit of each scheme was determined from the percentage of correctly identified keywords relative to linear amplification View full abstract»

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

    Publication Year: 1998 , Page(s): 489 - 492
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (396 KB)  

    Recently, several extensions of classical Shannon sampling theory to wavelet subspaces have been reported. This paper is devoted to uniform and periodic nonuniform oversampling in wavelet subspaces. Specifically, we provide a stability analysis and we introduce a technique for calculating the condition number of wavelet subspace sampling operators. It is shown that oversampling results in improved numerical stability. We consider the reconstruction from noisy samples and we characterize compactly supported scaling functions having compactly supported synthesis functions. Finally, it is shown that in the oversampled case the synthesis functions are not uniquely determined View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The discrete Gabor transform and the discrete Zak transform on a quincunx lattice

    Publication Year: 1998 , Page(s): 33 - 36
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (256 KB)  

    The discrete Gabor expansion on a quincunx lattice and its relation with the discrete Zak transform is presented. It is shown how the Zak transform can be helpful in determining Gabor's signal expansion coefficients and how it can be used in finding the dual window functions that correspond to a given elementary signal for this quincunx lattice. Furthermore, some examples are given and compared with Gabor's signal expansion on a rectangular lattice View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Modulated fractional Gaussian noise: a process with non-Gaussian wavelet details

    Publication Year: 1998 , Page(s): 361 - 364
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (288 KB)  

    The aim of this article is to propose a very simple model that is able to reproduce both scaling laws and scale dependent non-Gaussian wavelet details. It is based on the modulated version of a fractional Gaussian noise. Its power spectrum and the probability density functions of its discrete wavelet details are investigated and illustrated. Some examples based on simulated and real data are also given View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of classifier-optimal time-frequency distributions to speech analysis

    Publication Year: 1998 , Page(s): 585 - 588
    Cited by:  Papers (6)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (352 KB)  

    Discrete operator theory maps each discrete time signal to a multitude of time-frequency distributions, each uniquely specified by a kernel function. This kernel function selects some details to emphasize and other details to smooth. Traditionally, kernels are chosen to impart specific properties to the resulting distributions, such as satisfying the marginals or reducing cross-terms. Given a labeled set of data from several classes, we seek to generate a kernel function that emphasizes classification relevant details present in the distribution. In this paper, we extend our previous work on class dependent time-frequency distributions. Previously, the discriminant function did not consider the within-class to between-class variance of coefficients, and was vulnerable to choosing very “noisy” features View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A multifractal wavelet model for positive processes

    Publication Year: 1998 , Page(s): 341 - 344
    Cited by:  Papers (7)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (364 KB)  

    In this paper, we describe a new multiscale model for characterizing positive-valued and long-range dependent data. The model uses the Haar wavelet transform and puts a constraint on the wavelet coefficients to guarantee positivity, which results in a swift O(N) algorithm to synthesize N-point data sets. We elucidate our model's ability to capture the covariance structure of real data, study its multifractal properties, and derive a scheme for matching it to real data observations. We demonstrate the model's utility by applying it to network traffic synthesis. The flexibility and accuracy of the model and fitting procedure result in a close match to the real data statistics (variance-time plots) and queuing behaviour View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A channelized cross spectral method for improved frequency resolution

    Publication Year: 1998 , Page(s): 101 - 104
    Cited by:  Papers (1)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (320 KB)  

    A common problem in signal processing is the detection of weak tones in noise. Most tone recovery methods assume the presence of a single tone, or that the existing tones are sparse and well separated in frequency. In this paper, it is shown that tones may be detected and accurately estimated by first estimating a frequency error using the cross-spectral phase. The signal is then heterodyned by the error to translate it to the frequency of a Fourier basis element. The method has high gain, and may be used to isolate and recover tones which are not well separated View full abstract»

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