Uvi_Wave version 3.0
-=< WAVELET TOOLBOX for MATLAB >=-
------------------------------------
= Copyright =
--------------------------------------------------------------------
(c) Copyright 1996, by Universidad de Vigo under GNU conditions.
For any question about the toolbox files:
e-mail: Uvi_Wave@tsc.uvigo.es
ftp server:
ftp.tsc.uvigo.es
under /pub/Uvi_Wave/matlab directory
World Wide Web:
http://www.tsc.uvigo.es/~wavelets/uvi_wave.html
= What is the Wavelet Toolbox =
--------------------------------------------------------------------
This Wavelet Toolbox for Matlab intends to help the researches and
education professionals interested on any of the possibilities that
Wavelet Processing offers.
We hope the tools developed into a widely used environment as Matlab,
can be easily accepted by the scientific community.
In the next paragraphs, a short description of the main features of the
Uvi_Wave Toolbox can be found.
- Discrete Wavelet Transform: calculation and visualization of
the Discrete Wavelet Transform (DWT) and Inverse Discrete Wavelet
Transform (IDWT) for an unidimensional or bidimensional signal.
Routines to implement the typical filter bank structure have been
programmed.
- Scale Function and Wavelet Function: calculation and visualization of
the discrete approximation scale and wavelet functions, obtained as the
equivalent iterated filter that corresponds to the lower branch in the
filter bank structure.
- Multiresolution analysis: the objective of these routines
is the representation of the signal information into a set of
approximation signals and detail signals at different resolutions.
- Wavelet design: design of a set of filters that lead to the wavelet
decomposition and reconstruction. The selected design algorithms includes
orthogonal (Daubechies, Symlets, Battle-Lemarie, maximally flat, Remez) and
biorthogonal (Spline solution) families. Various filter regularity estimates
are also included.
- Wavelet Packet Transform: calculation of the Discrete Wavelet Packet
Transform for unidimensional or bidimensional signals. The routines allows
to set the filter bank structure, that is, the subband decomposition we want
to perform. Moreover, basis selection algorithms with different cost functions
are included.
- Scalogram: discrete-time simulation of the scalogram of a continous
signal using the Morlet wavelet as analyzing function.
In addition some demos and utilities for subband managing and viewing
have been included.
Despite the functions are clearly commented, we are working on the
documentation.
= Configuration of the Wavelet Toolbox =
--------------------------------------------------------------------
* The Subdirectory tree
"""""""""""""""""""""""""""""""
The Wavelet Toolbox's subdirectories are arranged as follows:
WAVELET_HOME
|
|_______ WVT
|_______ WPACKETS
|_______ WFILTER
|_______ WTUTILS
|_______ WPUTILS
|_______ SCAL
|_______ WDEMO
WAVELET_HOME Directory:
This is the root directory, that contains the STARTUP.M
file, a brief description of Uvi_Wave and the usage conditions.
WVT Directory:
The routines necessary to compute the wavelet transform
and to perform a multiresolution analysis can be found here.
WPACKETS Directory:
Contains all the wavelet packet transform functions,
together with some related utilities and the basis
selection algorithms. Some cost functions that can
be used with the selection routines are provided.
WFILTER Directory:
Contains the functions for generating the different filter
families, together with three regularity estimates.
WTUTILS Directory:
This contains all the viewing, alignment calculation,
band extraction/insertion/..., size finding and most
other utilities for the 1D-2D wavelet transform.
WPUTILS Directory:
Contains the viewing, band extraction/insertion/..., size
finding and miscellanous utilities for the 1D-2D wavelet
packet transform.
SCAL Directory:
Contains the functions to compute the sampled Morlet
wavelet, the scalogram of a sampled continous function and
a utility for 3D viewing of the scalogram.
WDEMO Directory:
Contains some demo programs and some data examples
(images and signals) used in the demos. A test image
generator also exists in this directory.
All Wavelet Toolbox's directories should be loaded into the
MATLABPATH variable to execute the toolbox's functions (maybe the
WDEMO can be left out). This task is made by the STARTUP.M file,
but surely this file should be edited to configure the directories
(at least the local WAVELET_HOME directory). See the next point
for help on configuring the toolbox.
* The STARTUP.M file
"""""""""""""""""""""""""""""""
What does Wavelet's STARTUP.M file do?
- Loads the four directory entries into MATLABPATH
- Sets a default method for wavelet subband alignment
Editing the STARTUP.M file:
The first thing you find is a global variable named
WAVELET_HOME_PATH, just before the "ABOUT THE WAVELET..." text.
There is a pre-installation check to avoid repeating the toolbox's
paths in the MATLABPATH variable.
Under this section there's a Copyright message display and
the definition of the WAVELET_HOME_PATH variable. Well, this is
what you must set to your Wavelet Toolbox's home directory.
If you preserve the original directory tree, then nothing
more is to do now. However, you may want to change any of the
subdirectories. Then make it under the previous declaration.
There's one string variable for every toolbox's directory. Change
them as you need.
The last section of STARTUP.M sets the WTCENTERMETHOD variable
to its default value of 0. If you want to start matlab with any
other value, change it here. See help on WTMETHOD from matlab for
allowed values of WTCENTERMETHOD. The alignment method can be
set to other value at anytime within the matlab session.