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

The Wavelet Digest - Volume 1, Issue 7 (October 23, 1992)


Today's Editor: Wim Sweldens

Today's Topics:
   1. Lost messages
   2. Book announcement
   3. Preprint available on Galerkin wavelet methods
   4. Question
   5. Question
   6. Papers available: The Multiresolution Fourier Transform
   7. Reference lists available via ftp

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.

Archive site:
    Anonymous ftp to maxwell.math.sc.edu (129.252.12.3),
    directory /pub/wavelet/archive.

Gopher server: bigcheese.math.sc.edu.

Current number of subscribers: 2179


From: Wavelet digest Date: Mon, 12 Oct 92 16:50:52 EDT Subject: Lost messages LS, We were informed that some mail sent to the wavelet user was lost during the last couple of weeks, in some cases without notifying the sender. If you sent a submission to the digest and it is not included, please resend your message. We are sorry for this inconvenience. Also, people who submit questions to the digest are invited to summerize and resubmit relevant answers. Thanks for your collaboration. Wim Sweldens, Bjorn Jawerth
From: Lora G. Weiss, Penn State Univ. Data: October 20, 1992 EST Subject: NEW BOOK ANNOUNCEMENT WAVELET THEORY AND ITS APPLICATIONS Randy K. Young Applied Research Lab Penn State Univ. (814) 863-4499 Kluwer Academic Publishers ISBN: 0-7923-9271-X ordering info: call: (617) 871-6300, ask for Customer Service, give ISBN (0-7923-9271-X) & VISA... or write: Kluwer Academic Publishers, PO Box 358, Accord Station, Hingham, MA 02018, with ISBN and check/money order/etc. %%%%%%%%%%%%%%%% CHAPTERS %%%%%%%%%%%%%%%%%%%%%%%%%%%% Chapter 1: Introduction/Background Chapter 2: The Wavelet Transform Chapter 3: Practical Resolution, Gain, and Processing Structures Chapter 4: Wavelet Theory Extensions and Ambiguity Functions Chapter 5: Linear Systems Modelling with Wavelet Theory Chapter 6: Wideband Scattering and Environmental Imaging This book reviews, extends, and applies wavelet theory. The book concentrates on the practical applications of wavelet theory. Many pictures (86 of them) provide visualizations of wavelet theory and its new extensions, as well as relationships to established concepts. Wavelet theory is integrated with other general theories, including linear systems theory and template matching or matched filtering. These relationships create analogies with related research and connections to practical applications. In addition, by demonstrating the effectiveness of wavelet theory in these general applications, many other specific applications may be improved. Temporal and spatial signals and systems are considered. The properties of the wavelet transform representation are sensitive to the chosen mother wavelet (the kernel of the wavelet transform, analogous to the exponential function in a Fourier transform). These properties are examined and techniques for analyzing these sensitivities are presented. Wavelet theory is extended with the new mother mapper operator that efficiently maps a wavelet transform with respect to one mother wavelet to a new wavelet transform with respect to a different mother wavelet. The mother mapper efficiently calculates concise wavelet representations that utilize multiple mother wavelets. The mother mapper operator is also employed to efficiently compute "cross" wavelet transforms or wideband cross ambiguity functions; these "cross" operators extract the "commonalities" between two signals or systems to determine the existence or structure of these commonalities. An original system model, the space-time-varying (STV) wavelet operator, is constructed with wavelet theory. As a special case, the STV model can represent linear time-invariant (LTI) systems. LTI systems are represented by the one-dimensional (1D)impulse response. This one-dimensional impulse response is the center slice of the two-dimensional (2D) STV representation. Both the LTI and STV models can be made to also vary with time (leading to 2D and 3D models, respectively). Physically, the STV wavelet operator creates an output by summing weighted, scaled and translated replicas of the input; these weights are the new system model. This is analogous to the LTI system model in which the output is a weighted sum of translated replicas of the input signal; with the weights being the LTI system model, the impulse response. Obviously, time scaling is the additional feature of the STV wavelet representation and is also the key to efficient representations of the wideband reflection or scattering process and improved estimation gains. By including the scaling operation as part of the STV system model (that is independent of time), the estimation process for this new system model can account for the linear time variation of the system and, thus, have a valid model over a longer interval of time. By estimating over a longer interval of time, more robust and higher gain and resolution estimates can be formed. Finally, the advantages of the new STV model are exploited to characterize or image an environment. Note from the editor: this book is available and the price is 67.50 US$.
From: Shann Wei-Chang, National Central University, Taiwan. Date: Sat, 17 Oct 92 16:20:37 -0800 Subject: Preprint available The following paper was done in May, 1991, with Prof. Jinchao Xu, when I was a graduate student in Penn State. Our point of view was from the finite element method. The key point is that the derivatives of the one dimensional piecewise linear hierarchical basis is exactly the Haar functions. Or, the lowest order of a family of Daubechies' wavelets with compact support. We thus extend this idea to wavelets with higher approximation power. The article was accepted by Numerische Mathematik. Jinchao Xu and Wei-Chang Shann Galerkin-wavelets methods for two-point boundary value problems Abstract. Anti-derivatives of wavelets are used for the numerical solution of differential equations. Optimal error estimates are obtained in the applications to two-point boundary value problems of second order. The orthogonal property of the wavelets is used to construct efficient iterative methods for the solution of the resultant linear algebraic systems. Numerical examples are given. Key words. Wavelets, finite element method, two-point boundary value problem, conjugate gradient method, preconditioner, piecewise linear hierarchical basis. A PostScript file with the text and graphics of this article can be obtained from the anonymous ftp account in dongpo.math.ncu.edu.tw (140.115.25.3). The file is pub/shann/wavelets/xu.ps. This is an ASCII file of length about 422K. You can take the compressed version xu.ps.Z if you know how to do it. Printed copies can be requested from the address below. Wei-Chang Shann, Associate Professor Dept of Mathematics, National Central University, Chung-Li, Taiwan, R.O.C. (03)425-6704 ext. 166 (work); (03)425-7379 (work.FAX) shann@math.ncu.edu.tw
From: Keith Deacon Date: Mon, 12 Oct 92 16:15:46 MDT Subject: Question Has anybody done any work with moving, i.e. Lagrangian, wavelets in the solution of physical problems like advection-diffusion? It is well known that orthogonal wavelets and shape functions on compact support retain their orthogonality for certain finite distance translations. For translations between allowable grid points orthogonality is lost and coefficients vary dramatically. When wavelets are used to represent certain physical phenomena their coefficients can be related to grid nodes through Galerkin or collocation techniques to solve Eulerian formulated equations. The physical problems of convection and diffusion involve moving concentrations or mass. While the equations can be formulated on a fixed grid, the Eulerian grid, the equations can also be formulated from the viewpoint of the moving mass, Lagrangian equations. Is there a way to take advantage of the multi- resolution properties of wavelets in the Lagrangian frame of reference? This would probably require the formulation of moving multi-resolution wavelets and shape functions. Any references or correspondence would be welcome. Thanks, Replies can be emailed to kdeacon@nmsu.edu or kdeacon@wsmr-emh74.army.mil Phone correspondence to myself or Mr. Ron Meyers at (505) 678-4037 Thank you again, Keith Deacon
From: Bhavik Bakshi. Massachusetts Institute of Technology. Date: Mon, 12 Oct 92 16:32:47 EDT Subject: Question Hi, My question is regarding multi-dimensional wavelets. The usual way of generating wavelets in higher dimensions is by taking a tensor product. For an n-dimensional system, this gives (2^n - 1) wavelets, and one scaling function. Thus, the number of wavelets increases exponentially, as the dimension increases. Is there any way of generating multi-dimensional wavelets where the number of wavelets does not increase as an exponential power of the number of dimensions? Thank-you. Bhavik Bakshi Department of Chemical Engineering Massachusetts Institute of Technology.
From: Andy Davies, University of Warwick, Date: Mon, 12 Oct 1992 15:01:16 +0100 Subject: Papers available: The Multiresolution Fourier Transform The Multiresolution Fourier Transform ------------------------------------- A generalized form of Wavelet Transform which enables adaptive analysis of the `time-frequency' plane over multiple scales has been developed by the Signal Processing Group at the University of Warwick, England. The MFT has been applied successfully to segmentation problems in both image and audio signal analysis. Details of these and properties of the transform can be found in the following: A Generalized Wavelet Transform for Fourier Analysis : the Multiresolution Fourier Transform and its Application to Image and Audio Signal Analysis R. Wilson, A.D. Calway, and E.R.S. Pearson. IEEE Trans. Information Theory, Special Issue on Wavelet Transforms and Multiresolution Signal Analysis, vol. 38, 2, 1992. Abstract A wavelet transform specifically designed for Fourier analysis at multiple scales is described and shown capable of providing a { em local representation which is particularly well suited to segmentation problems. It is shown that by an appropriate choice of analysis window and sampling intervals, it is possible to obtain a Fourier representation which can be computed efficiently and overcomes the limitations of using a fixed scale of window, yet by virtue of its symmetry properties allows simple estimation of such fundamental signal parameters as instantaneous frequency and onset time/position. The transform is applied to the segmentation of both image and audio signals, demonstrating its power to deal with signal events which are localised in either time/space or frequency. Feature extraction and segmentation are tackled through the introduction of a class of multiresolution Markov models, whose parameters represent the signal events underlying the segmentation. In the case of images, this provides a unified and computationally efficient approach to boundary curve segmentation; in audio analysis, it provides an effective way of note segmentation, giving accurate estimates of onset time and pitch in polyphonic musical signals. Additional information can also be found in: R. Wilson and A.D. Calway, ``A general multiresolution signal descriptor and its application to image analysis'', Proc. EUSIPCO-88, pp. 663-666, Grenoble, 1988. A. Calway and R. Wilson, ``A unified approach to feature extraction based on an invertible linear transform'', Proc. 3rd Int. Conf. Image process. & its Appl. , pp. 651-655, 1989. A.D. Calway, The Multiresolution Fourier Transform: A General Purpose Tool for Image Analysis , Ph.D. Thesis, Warwick Univ., 1989. A.D. Calway and R. Wilson, ``Curve extraction in images using the multiresolution Fourier transform'', Proc. IEEE ICASSP-90 , pp. 2129-2132, Albuquerque, 1990. E.R.S. Pearson and R. Wilson, ``Musical event detection from audio signals within a multiresolution framework'', Proc. ICMC-90, Glasgow,1990. E.R.S. Pearson, The Multiresolution Fourier Transform and its Application to the Analysis of Polyphonic Music , Ph.D. Thesis, Warwick Univ., 1991. A.R. Davies and R. Wilson, ``Curve and corner extraction using the multiresolution Fourier transform'', Proc. IEE Conf. on Image Process. and its Appl. , Maastricht, Holland, 1992. R. Wilson, A.D. Calway, E.R.S. Pearson and A.R. Davies. ``An introduction to the multiresolution Fourier transform and its applications'', Res. Report RR204, Dept. of Computer Science, University of Warwick, 1992. Tao-I Hsu, A.D. Calway and R.Wilson, ``Analysis of structured texture using the multiresolution Fourier transform'', Res. Report RR226, Dept. of Computer Science, University of Warwick, 1992. A.D. Calway, H.Knutsson, and R. Wilson. ``Multiresolution estimation of 2-d disparity using a frequency domain approach'', Proc. British Machine Vision Conf., Leeds, September 1992. For copies of the above and further information contact: Dr Roland Wilson Department of Computer Science University of Warwick Coventry CV4 7AL England email: rgw@dcs.warwick.ac.uk
From: Wavelet digest Date: Mon, 12 Oct 92 18:50:32 EDT Subject: Reference lists available via ftp The reference lists sent to the digest are now also available via anonymous ftp. There are three reference lists from Milos Doroslovacki (~100 entries), Christopher Koenigsburg (~80 entries) and Tom Koornwinder (~280 entries). We would like to thank these people for their contributions. To retrieve them, connect to maxwell.math.sc.edu and look for /pub/wavelet/references/*.ref. Obviously there is quite some overlap between those lists. If there is anybody who would be interested in merging these lists into one with a consistent format (such as bibtex), please get in touch with the wavelet digest.

<webmaster@wavelet.org>