July 2015: new version 3.0!
This version fixes two bugs in the curvelet transform (size issue depending on the size of the input image + wrong filters were generated for the last scale annulus in the curvelet option 2) + implementation of the curvelet option 3 (detect scales per each angular sectors)
In this toolbox, we implement the Empirical Wavelet Transform for 1D and 2D signals/images. The principle consists in detecting Fourier supports on which Littlewood-Paley like wavelets are build. In 2D, we revisit different well-known transforms: tensor wavelets, Littlewood-Paley wavelets, ridgelets and curvelets.
The toolbox also provides the scripts used to generate the experiments in the papers:
- J.Gilles, "Empirical wavelet transform" to appear in IEEE Trans. Signal Processing, 2013.
Preprint available at ftp://ftp.math.ucla.edu/pub/camreport/cam13-33.pdf
- J.Gilles, G.Tran, S.Osher "2D Empirical transforms. Wavelets, Ridgelets and Curvelets Revisited", SIAM Journal on Imaging Sciences, Vol.7, No.1, 157--186, 2014.
Preprint available at ftp://ftp.math.ucla.edu/pub/camreport/cam13-35.pdf
- J.Gilles, K.Heal, "A parameterless scale-space approach to find meaningful modes in histograms - Application to image and spectrum segmentation", submitted 2014.
Preprint available at
See the README file inside the archive for more instructions