SCILAB
WAVELET
TOOLBOX

SWT candidate 6 windows version released for Scilab 4.x and 5.x finally!

About

Wavelet is a powerful signal processing tool developed and developing in the last two decades. And Scilab is a freely redistributable scientific computing software quite like Matlab and Octave, from French from French National Institute of Automation and Information INRIA. Scilab Wavelet Toolbox is a free software package to enable you using g wavelet analysis tools Freely in Scilab on most OSes including GNU/Linux, BSD and Windows.

Scilab Wavelet Toolbox is designed to work with any Scilab Image Processing Toolbox like SIP or SIVP for displaying 2-D results. Then one could use ImageMagick to do post-processing for websites, presentations or papers.

Scilab Wavelet Toolbox is developed in C and referenced to GNU Scientific Library GSL and WaveLab from Stanford University. There are two scilab version WaveLab from Chinese Xi'an JiaoTong University and a French University. The reason we still try to implement a C toolbox mostly compatible with Matlab is that most algorithm developers are most familiar with Matlab Wavelet Toolbox. And obtaining scilab C interface experience is important for future toolbox development, indepenent with scilab internal function changing.

Scilab Wavelet Toolbox is born for use.: It is intended for education courses, signal processing research and specific interests. For example, you could use this toolbox to help your courses, do ultrasonic signal and seismic signal analysis, fuse different images for medical diagnosis or something for fun. Anyone who familiar with Matlab Wavelet Toolbox could easily get along with it. Moreover, new comers would find helpful information from Matlab books and do what they want on Scilab.
Scilab Wavelet Toolbox is completely free to the public.: We hope that would help reducing illegal using of pirate version of Matlab. For those who can not afford buying Matlab in developing countries, we provide you an alternative way to do your research.
Scilab Wavelet Toolbox is simple.: This toolbox is to mimic Matlab Wavelet Toolbox and so we think it is easy to get started. Somehow, We do not provide high abstract level functions like wavelet denoising because it is your creativity to achieve that. Furthermore, we left the high level functions to your imagination for your great work.
Scilab Wavelet Toolbox is fun.: Many students are overwhelmed by mathematics of the textbooks. With this great software of Scilab, computation could be easily done and the theories could be easily verified within minutes. Our toolbox is also intended for application developers. And we are sure you could enjoy your project with our wavelet toolbox.

What Scilab Wavelet Toolbox supposed to do

Discrete Fast Wavelet Transform, daubechies wavelets
1-D single level signal decomposition and reconstruction
1-D multi-level signal decomposition and reconstruction
2-D single level image decomposition and reconstruction
2-D multi-level image decomposition and reconstruction
3-D single level signal decomposition and reconstruction
stationary wavelet transform
continuous wavelet transform
dualtree real and complex wavelet transform

Screenshot

Screenshot 1 is the ultrasonic signal wavelet decomposition result while screenshot 2 is the classical Lena image wavelet decomposition.

Screenshot1 Screenshot2

Licenses: GPL

Current Version: alpha

Development Group:

Professor Mei, Supervisor
Roger Liu (rogeruclan at gmail dot com), Software Developer
Isaac Zhi (QFWFQzhi at gmail dot com), Software Developer
Jason Huang, Software Tester
Du HuiQian, Application Developer

All from Department of Electrical Engineering, Beijing Institute of Technology, P.R.C.

News

March 11th, 2009

With almost one year delay, finally candidate 6 is released. Originally I was going to attend a PhD course about wavelets in TelecomParis, ENST. However, it is totally in French and I cannot do the assigned projects and attend the exam in French. So it is too much for me.

My current situation: I had a deep trouble in my current project. I may drop out. But I will continue this project

One surprise from Trinidad & Tobago. One guy from there asked me about SWT. So I am amazed Scilab reaches every corner of the world, especially during this period of finacial crisis.

May 20th

I am registered in Paris VI university, pierre et marie curie. So I am a PhD student now in computer graphics. My current project is for building modeling. And especially I am using my toolbox for facade analysis. Wiki page is open.

Jan 8th

I got permission from Professor Ivan Selesnick so I could convert his code into SWT, complex wavelet transform. He is in Polytechnic University, Brooklyn, NY, USA. Please go to his webpage for more information on dualtree complex wavelet transform.

Nov 5th

After 12 hours flight, I arrived in Paris. I am now an employee of INRIA.

June 17th

*NIX version released! Next release will be 0.1.0 with some WaveLab and Time-Freqency Functions intergrated!

June 16th

RC4 test passed on Windows and FreeBSD. Windows version install package is provided. *NIX version will be released soon. Users should also install SIVP or SIP to load and display image and transform results. In this release, demos and help topics are provided.

June 4th

The author finally solve the discrete wavelet transform procedure on rc4. During normal analysis process, wkeep must be performed before dyaddown to reduce errors. Current version is pre-alpha windows version in tolerance of fault input. However, the warning message is not meaningful. For the final version, warning message will be talored to give user more information.

Nov.30th

Roger Liu and Du HuiQian visited INRIA SCILAB Team.

Oct. 8th

Release Delayed due to heavy travel arrangment. The author has too tight travel in Sep. and will go to U.S. in Oct. and may go to France in late Nov. Sorry for the delay. But the development is still ongoing. The internal release has been done in Sep. 05 but bugs not removed. This work has been recognized by Scilab Contest and Workshop 2006 in China ZheJiang University as the first prise.

July 9th

The fourth release candidate development is ongoing as an author coming back from job changing. Also a phD student joined us as a software tester. Welcome Jason Huang on board! The estimated release date would be before September. In this release, coiflets, symlets and spline biorthorgnal wavelets are supported. And function usage would be more compatible with Matlab Wavelet Toolbox. See you in September!

Mar. 25th

Third candidate version released to fix inverse transform bugs. When image has length and width not equal, the backward transform result is not correct on rc2. Now the bug is fixed.

Lena image under data directory is removed for sake of smaller package size. Please download SIVP and check the Lena image in its package.

There is no need to use gmake instead of make under BSD now. 'make' runs smoothly under all POSIX compliant OSes.

Also an introduction slide is made on sourceforge downloading site.

Mar. 16th

Second candidate version released to fix a critical bug. On rc1, transform on Images with unequal length and width will have strange result. On rc2, this bug is fixed.

Mar. 11th

Windows port has been made. Website updated.

Mar. 1st

First candidate version 0.1.0rc released. Scilab also adds our toolbox into their website on contribution page under the category of math. Bug reports, Comments, suggestions and feelings are highly appreciated and should be sent to us through above email address.

Feb. 1st

Homepage updated. SWT Logo added. With Chinese Lunar New Year in progress, we slow down the releasing progress. After the 7-day holiday, we will release the software package with automake and autoconfig utilities. Until now, we are using a simple builder.sce script to build the library. In addition, the documentation is also in progress. We hope to make the first pre-alpha release at the end of February.

Jan. 19th

We have developed a couple of C functions interfacing with Scilab based on Scilab Interface Examples, including one dimensional analysis and synthesis functions and two dimensional counterparts. Moreover, we have made initial experiments with both ultrasonic signals and images. The result is quite exciting. In addition, both FreeBSD and Debian Linux work along with our toolbox. Therefore we believe we have set up a good baseline to go forward.

And we would place the original C functions on the sourceforge website within a month to let the public have a glance at it. In order to make this toolbox to deal with error inputs, and also from software engineering point of view, the whole toolbox should be rewritten from this point.

Downloading

Downloading site is at Scilab Wavelet Toolbox on Sourceforge

GNU/Linux and BSD

Please check the enclosed text files to install the package. Uncompress the package, say swt-0.1.0rc3.tar.gz, using tar utility. Then change the directory to the untared directory. Type "./configure", "make" and "make install". If you want to uninstall swt, type "make uninstall".

Windows

We have not made an automatic install package. Just unzip the package in Scilab contrib directory. For example, if you have install your Scilab software under Program Files, the place you would mostly like to place swt is C:/Program Files/scilab/contrib.

Load the toolbox

Start Scilab, click toolbox menu at the upper right corner of Scilab GUI Window, on item swt.

Documentation

We have a reference sheet and an introduction slide published on the sourceforge along with the software release. For detailed help information, please check the help chapter in the scilab GUI. In addition, now we are working on an userguide. But this progress is quite slow. Please check Scilab Wavelet Toolbox on Sourceforge.