[Home] [Back Issues] [Search] [Vol 1] [Vol 2] [Vol 3] [Vol 4] [Vol 5]

The Wavelet Digest - Volume 3, Issue 5 (March 18, 1994)


Today's Editor: Wim Sweldens
                sweldens@math.sc.edu


Today's Topics:
    1. Answer:   WD 3.4 (#1): Meyer's books
    2. Answer:   WD 3.4 (#1): English translation of the book by Yves Meyer
    3. Meeting:  15th Franco-Belgian Meeting on wavelets and statitics
    4. Announce: Wavelets on the World Wide Web
    5. Preprint: Introductory article on the wavelet packet transform
    6. Question: New wavelet ?
    7. Answer:   WD 3.4 (#2): Wavelets and time series
    8. Job:      Postdoc Position available at CCETT in Rennes
    9. Question: Wavelets and speech classification
   10. Preprint: Noise suppression and signal compression using ...
   11. Contents: Numerical Algorithms, Volume 6, No. 3 - 4,
   12. Contents: Constructive Approximation Vol.10 Number 1
   13. Question: Multipole computations

Submissions for Wavelet Digest: 
    E-mail to wavelet@math.sc.edu with "submit" as subject.

Subscriptions for Wavelet Digest: 
    E-mail to wavelet@math.sc.edu with "subscribe" as subject.
    To unsubscribe, e-mail with "unsubscribe" followed by your e-mail
    address as subject. To change address, unsubscribe and resubscribe.

Archive site, preprints, references and back issues:
    Anonymous ftp to maxwell.math.sc.edu (129.252.12.3),
    directory /pub/wavelet.

Gopher server: gopher.math.sc.edu.

Current number of subscribers: 3255


From: folland@math.washington.edu Subject: Answer: Meyer's books (WD 3.4 #1) A comment on topic #1 in the latest wavelet digest (English translation of Meyer's book): Meyer's book consists of three volumes. (The last one, jointly written with Coifman, has little to do with wavelets.) The first volume makes some references to the second volume, and in particular uses some results from the second volume on Calderon-Zygmund operators. Now, the English translation is only of the first volume. On the one hand, the references to the second volume were left in the translation. On the other hand, there's no indication that the second volume is being translated too. I am puzzled by this, and other readers may be too. Gerald Folland [Answer]
From: Vern Paxson <vern@ee.lbl.gov> Subject: Answer: English translation of the book by Yves Meyer (WD 3.4 #1) > Note from the editor: > Yves Meyer, Wavelets and Operators, Cambridge University Press, 1992 > ISBN 0 521 42000 8, hardback, 223 pp. Perhaps Bhaskar is asking about another book by Yves Meyer: Wavelets: Algorithms & Applications Yves Meyer, Translated and Revised by Robert D. Ryan, SIAM, 1993, ISBN 0-89871-309-9 Vern Paxson vern@ee.lbl.gov Information and Computing Sciences ucbvax!ee.lbl.gov!vern Lawrence Berkeley Laboratory (510) 486-7504
From: Anestis.Antoniadis@imag.fr Subject: Meeting: 15th Franco-Belgian Meeting on wavelets and statitics Announcing A Conference (15th Franco-Belgian Meeting of Statisticians) on ``Wavelets and Statistics'', to be held from November 16 to November 18, 1994, in Grenoble, France. Organized by the Permanent Committee of the ``Rencontres Franco-Belges des Statisticiens'' Wavelet theory has proved to be a powerful tool in many areas of statistics. This conference will bring together researchers or practitioners, analysts or statisticians, to present an up-to-date review of the current state-of-the-art, to assess the relative advantages and limitations of wavelets for various statistical applications and to identify trends for future development. Among the topics planned to be represented at the conference are the applications of wavelets to Signal and Time Series analysis, Nonparametric Statistics and Neural networks, Random Fields and Image Analysis. In each of these fields there are problem areas in which wavelet analysis has proved to be useful, if not indispensable. The interdisciplinary character of the subject will be particularly emphasized in this conference by the importance given to application oriented aspects and by contributions incorporating co-operation between theory and practice. Location: The meeting will take place in the Grand Hotel de Paris, located in Villard de Lans, in the heart of the Vercors, near Grenoble. Villard de Lans is accessible by public transportation from Grenoble. The distance from Grenoble is approximately 34 Km. Local Organising Committee: Prof A. Antoniadis IMAG-LMC, Laboratoire de Statistique et Modelisation Stochastique University Joseph Fourier, Grenoble, France Partial List of Invited Speakers: B. Delyon (France), A. Cohen (France), D. Picard (France), B. Silverman (Great Britain), G. Nason (Great Britain), J. Istas (France), G. Oppenheim (France), K. Tribouley (France), Q. Zhang (France). Correspondence and Further Information: Those interested in participating, contributing or receiving future announcements, please contact the organisers: Anestis Antoniadis RFBS, IMAG - LMC Equipe de Statistique. Universite Joseph Fourier B.P. 53 38041 Grenoble Cedex 09 France Fax: (33)76 631263 Phone: (33)76 514306 E-mail: antonia@imag.fr
From: Andreas Uhl, University of Salzburg, Austria Subject: Announce: Wavelets on the World Wide Web Wavelets on the World Wide Web ? Are there any pages on wavelets in the WWW ? I maintain a small page containing useful ftp-sites, a bibliography, ...... (www.mat.sbg.ac.at/ ~uhl/wav.html). I'd like to invite everybody concerned with wavelet to create his own page - please let me know ! Uhl Andreas Univ. Salzburg uhl@edvz.sbg.ac.at
From: Mac A. Cody <mcody@aol.com> Subject: Preprint: Introductory article on the wavelet packet transform "The Wavelet Packet Transform" by Mac A. Cody will appear in the April, 1994 issue of Dr. Dobb's Journal (DDJ) magazine (Miller Freeman, Inc.; ISSN 1044-789X). This article is a more "accessible" treatment of the topic as compared to most preprints and transaction papers. It is the third in a series that has been published in this magazine by the author ("The Fast Wavelet Transform", DDJ April, 1992 and "A Wavelet Analyzer", DDJ April, 1993). As with the previous articles, theory has been minimized in favor of the practical aspects of how the transform works and its implementation. The wavelet packet transform is approached by building from its roots in the fast wavelet transform. The transform's structure is based upon a binary tree; each node representing the applica- tion of a decimating, quadriture mirror filter (QMF) pair to a signal. Different wavelet packet bases can be created by selec- tively "pruning" the binary tree. The pruning also generates unique partitioning of the original signal bandwidth into a dis- joint cover of frequency subbands. The wavelet packet transform is, in effect, an arbitrary adaptive tree-structured filter bank. ANSI C implementations of the wavelet packet transform, the inverse transform, and supporting data structures are presented and explained. An IBM PC program that demonstrates the operation of the wavelet packet transform has been written by the author. The program is available through Dr. Dobb's Journal via annonymous ftp from site ftp.mv.com (192.80.84.1) in the /pub/ddj directory.
From: Richard K Barrett <richard@astro.gla.ac.uk> Subject: Question: New wavelet ? Dear Waveleteers, I have recently been introduced to the mystic art of wavelet transforms thanks to the wavelets section in the new edition of Numerical Recipes. As a complete novice, I have a question for the more experienced reader of the Wavelet Digest. I apologize in advance for the length of this question, but I don't know how to make it short *and* intelligible. In Numerical Recipes it says that the wavelet filter, H(w) (w=frequency), which is the fourier transform of the sequence of wavelet filter coefficients, c(1), c(2), ..., c(N), must satisfy : H(w) is periodic, with period 2*pi, (1) |H(0)|^2 = 2 (|H| denotes the modulus of H), (2) |H(w)|^2 + |H(w+pi)|^2 = 2 (3) and must satisfy an `approximation condition of order p', which corresponds to H, and its first p-1 derivatives are zero at w = pi. (4) I can derive a wavelet filter satisfying an approximation condition of infinite order (`p=infinity') using the famous infinitely differentiable but not analytic function given by E(x) = exp{1/(x^2 - 1)}, for -1<x<1, and E(x)=0 otherwise, which is only non-zero for -1<x<1. A simple linear transformation of the x co-ordinate can give a function with a `hump' between any two points a and b, say, and the value zero elsewhere. Therefore define E(a,b,s;x) = exp{s^2/[(x-a)(x-b)]}, for a<x<b, and E(a,b,s;x) = 0 otherwise. (Here s is simply a (non-zero) parameter, which may, for the moment, be taken to be 1.) Note that all derivatives of E with respect to x exist and are continuous, and all derivatives are zero for x<=a and x>=b. It is this last property that gives an approximation condition of infinite order. It is (fairly) easy to verify that the following definition of H(w) satisfies (1), (2), (3) and (4) with p=infinity : Put H(n,a,b,s;w) = sqrt(2) Cos {(2n+1)F(a,b,s;w)}, where n is any integer and a, b, and s are the parameters introduced above, and we require 0<=a<b<=pi (to simplify things, imagine that a=0, b=pi, s=1, and n=0). It is the form of the function F(a,b,s;w) that ensures (4) holds with p=infinity. It can be seen that if all derivatives of F (with respect to w) are zero at w=0 and w=pi then H(w) will have all derivatives zero at w=pi (note the role of (3) here). Furthermore, forcing F(w=pi)=pi/2 guarantees that H(w=pi) = 0, giving (4) with p=infinity. In terms of this function F, conditions (1) - (4) become, therefore : although not essential, we may assume that F is 2*pi periodic in w, (1') F(w=0) = 0, (2') F(w+pi) = (pi/2) - F(w) for 0<=w<=pi, (3') F(w=pi)=pi/2, all derivatives of F at w=0 and w=pi are zero. (4') In addition to these conditions, we might as well demand that F is infinitely differentiable w.r.t. w (this will ensure that, while the wavelet is not compact, the filter coefficients, c(i), will fall of faster than any polynomial). One solution to these conditions is the function satisfying the differential equation ( ' = d/dw) F'(a,b,s;w) = k*E(a,b,s;w) for 0<=w<=pi and F'(a,b,s;w) = -k*E(a,b,s;w) for pi<=w<=2*pi with initial condition F(w=0)=0, and where k is a renormalization constant chosen to ensure F(w=pi) = pi/2. The properties of these wavelets are (stop me if I'm boring you) : 1) In general the filter coeffs. compare well with the Lemarie wavelet. s may be used to control the rate at which the filter coefficients fall off. 2) The actual wavelets are smooth for s=8 and get `sharper' as s decreases. 3) For some parameter values the wavelet functions have a fractal appearance (some of them actually look kind of `hairy'), although I suspect that this is due to numerical error in the calculation of the wavelet coefficients and the inverse wavelet transform. One further point worth noting is that it is possible to multiply the function F by another function, g(w) say, and still get a wavelet with the same properties -- the only constraint is that we must have g(b)=1, and the derivatives of g(w) must all be zero at w=b to ensure that the derivatives of g*F are all zero there. At last, after all this nonsense, I arrive at my question. Is the wavelet derived a: interesting b: original I look forward to your replies. Richard Barrett. P.S. Wim Sweldens has told me that this wavelet is related to the Meyer Wavelet. I would like to know what the practical applications of such wavelets are.
From: Lonnie Hudgings <lonnie@golem.ps.uci.edu> Subject: Answer: Wavelets and time series (WD 3.4 #2) Dear Esben, In response to your inquiry regarding wavelet spectra for non- stationary time series, I would like to offer a paper of mine: "Wavelet Transforms and Spectral Estimation," by Lonnie H Hudgins. This paper, which became part of my Ph.D. dissertation in 1992, explores the relationships between wavelet spectra and Fourier spectra. Below is an abstract: "The wavelet transform is defined here in terms of the scale number variable s instead of the more common scale length variable a where a = 1/s. This convention permits the scale number for wavelets to compare directly with wave number in Fourier, and the theory of wavelet spectral estimation proceeds in parallel with that of Fourier transforms. After reviewing some motivation for the usual Fourier power spectrum, definitions are given for both one-sided and two-sided wavelet power spectra. These 'spectral densities' are then interpreted in the context of standard Fourier theory and banks of spectral estimation filters. Polarization then yields definitions for wavelet cross spectrum, wavelet cospectrum, and wavelet quadrature spectrum, which agree with their Fourier counterparts whenever we could reasonably expect them to. These tools also provide us with the helpful notions of wavelet transfer function, and wavelet coherence. Finally, the theory leads to a bivariate wavelet cross transform: a useful new function for pairs of signals which behaves like the Fourier cross transform, i.e., the magnitude emphasizes signal similarities, and the phase contains relative timing information between the two signals, while inheriting the 'zoom-in' properties of wavelets." The ideas developed in this paper have been successfully applied to detect, identify, and locate intermittent large scale coherent structures in turbulent fluid flow by analyzing the off-diagonal components in the generalized Reynolds stress tensor (see Physical Review Letters, Nov 1993). Good Luck, Lonnie
From: C. Guillemot, Christine.Guillemot@ccett.fr Subject: Job: Postdoc Position available at CCETT in Rennes Dear Colleagues, There is a postdoc position open in CCETT in Rennes (France) (joint research center on telecommunication and broadcasting) to work on joint wavelet filters/quantization optimization and more specially on techniques that can adapt locally to the spatial characteristics of images for compression purposes. For more information please contact Christine Guillemot CCETT, dpt TAV/CEP 4, rue du Clos Courtel 35512 Cesson-Sevigne France Phone: 33 99 12 45 04 fax: 33 12 40 98 e-mail: guillemo@ccett.fr
From: kripps@arjun.csc.tifr.res.in (A. Kirupairaj) Subject: Question: wavelets and speech classification Hello there, I'm working on the application of wavelet transforms to classification problems, particularly to finer classification of plosives in speech signal. Is there anybody with a similar interest, I would like to interact. I would also welcome any literature in related applications. Thanking you A. Kirupairaj Computer Systems and Communications group TIFR, Bombay Email: kripps@arjun.csc.tifr.res.in
From: saito@sdr.slb.com (Naoki Saito) Subject: Preprint: Noise suppression and signal compression using ... Preprint and reprint are available The following preprint is available via an anonymous ftp to pascal.math.yale.edu and the filename is "/pub/wavelets/snssc.ps.z". This is a compressed postscript file by gzip, so you need gunzip program to uncompress it. Title: Simultaneous Noise Suppression and Signal Compression using a Library of Orthonormal Bases and the Minimum Description Length Criterion} Author: Naoki Saito Abstract: We describe an algorithm to estimate a discrete signal from its noisy observation, using a library of orthonormal bases (consisting of various wavelets, wavelet packets, and local trigonometric bases) and the information-theoretic criterion called minimum description length (MDL). The key to effective random noise suppression is that the signal component in the data may be represented efficiently by one or more of the bases in the library, whereas the noise component cannot be represented efficiently by any basis in the library. The MDL criterion gives the best compromise between the fidelity of the estimation result to the data (noise suppression) and the efficiency of the representation of the estimated signal (signal compression): it selects the ``best'' basis and the ``best'' number of terms to be retained out of various bases in the library in an objective manner. Because of the use of the MDL criterion, our algorithm is free from any parameter setting or subjective judgments. This method has been applied usefully to various geophysical datasets containing many transient features. Also, the following REPRINT is available upon request via email or postcard to Naoki Saito. Title: Multiresolution Representations Using The Auto-Correlation Functions of Compactly Supported Wavelets Author: Naoki Saito and Gregory Beylkin Journal: IEEE Trans. Signal Proc., special issue on wavelets and signal processing, Vol. 41, No. 12, pp.3584-3590, 1993. Abstract: We propose a shift-invariant multiresolution representation of signals or images using dilations and translations of the auto-correlation functions of compactly supported wavelets. Although these functions do not form an orthonormal basis, their properties make them useful for signal and image analysis. Unlike wavelet-based orthonormal representations, our representation has (1) symmetric analyzing functions, (2) shift-invariance, (3) associated iterative interpolation schemes, (4) a simple algorithm for finding the locations of the multiscale edges as zero-crossings. We also develop a non-iterative method for reconstructing signals from their zero-crossings (and slopes at these zero-crossings) in our representation. This method reduces the reconstruction problem to that of solving a system of linear algebraic equations. Naoki Saito Schlumberger-Doll Research and Dept. of Mathematics, Yale University Email: saito@ridgefield.sdr.slb.com
From: publish@baltzer.nl (Daniel Baltzer) Subject: Contents: Numerical Algorithms, Volume 6, No. 3 - 4, Numerical Algorithms, Volume 6, No. 3 - 4, 1994, ISSN 1017 1398 Editor-in-Chief: Claude Brezinski pp 205-227: S. Elhay and J. Kautsky, Jacobi matrices for measures modified by a rational factor pp 229-244: A. Sidi, Convergence of intermediate rows of minimal polynomial and reduced rank extrapolation tables pp 245-273: J.C. Dodu, T. Eve and M. Minoux, Implementation of a proximal algorithm for linearly constrained nonsmooth optimalization problems and computational results pp 275-296: D. Amitai, A. Averbuch, S. Itzikowitz and E. Turkel, Asynchronous and corrected-asynchronous finite differences solutions of PDEs on MIMD multiprocessors pp 297-316: G. Plonka, Optimal shift parameters for periodic spline interpolation pp 317-351: A.-M. Bellido, Construction of iteration functions for the simultaneous computation of the solutions of equations and algebraic systems pp 353-378: M.K. Ng and R.H. Chan, Fast iterative methods for least squares estimations pp 379-418: J.-C. Fiorot, P. Jeannin and S. Taleb, New control massic polygon of a B-rational curve resulting from a homographic change of parameter pp 419-423: Book reviews
From: Prof. E.B. Saff <esaff@gauss.math.usf.edu> Subject: Contents: Constructive Approximation Vol.10 Number 1 CONSTRUCTIVE APPROXIMATION Contents Volume 10 Numbers 1 1994 1 Constructive Approximations to the Invariant Densities of Higher-Dimensional Transformations A. Boyarsky, P. Gora, and Y. S. Lou 15 On the Darling-Mandelbrot Probability Density and the Zeros of Some Incomplete Gamma Functions John S. Lew 31 Convex Polynomial and Spline Approximation in C[-1,1] Yingkang Hu, Dany Leviatan, and Xiang Ming Yu 65 Fourier Series of Functions Whose Hankel Transform is Supported on [0,1] Juan L. Varona 77 Best Uniform Approximation by Harmonic Functions on Subsets of Riemannian Manifolds P. M. Gauthier and D. Zwick 87 Simultaneous Lagrange Interpolating Approximation Need Not Always Be Convergent S. P. Zhou 95 Strong Converse Inequality for Kantorovich Polynomials W. Chen and Z. Ditzian 107 Lehmer Pairs of Zeros, the de Bruijn-Newman Constant \Lambda, and the Riemann Hypothesis George Csordas, Wayne Smith, and Richard S. Varga 131 Maximal Polynomial Subordination to Univalent Functions in the Unit Disk Vladimir V. Andrievskii and Stephan Ruscheweyh
From: Rudolph.Lorentz@gmd.de (Rudolph Lorentz) Subject: Question: Multipole computations I am interested in multipole computations. 1) Is there a source for Fortran or C programs which implement the approach of Greengard and Rokhlin? 2) Is there an approach using wavelets? If there is, I am interested in both references to the literature as well as sources for programs. Rudolph A. Lorentz Gesellschaft fuer Mathematik und Datenverarbeitung 53757 St. Augustin Germany E-mail: lorentz@gmd.de

<webmaster@wavelet.org>