110 likes | 132 Views
This project aims to implement discrete wavelet transformation and inverse transformation wrappers in Matlab to handle multi-dimensional data, such as 2D images, 3D turbulence data, or multi-attribute sensor readings. The project also explores possible applications of wavelets, including lossy compression, de-noising for images, and self-similarity analysis. The project further studies the phases of wavelet filters and approximates the delay amount using DSP methods. Bayesian methods may also be explored for missing detail modeling. The project concludes with a DSP perspective on problems related to non-zero phase filtering.
E N D
Experimenting with Multi-dimensional Wavelet Transformations Tarık Arıcı and Buğra Gedik
Outline of Project Goals • Writing discrete wavelet transformation and inverse transformation wrappers (in Matlab) to handle multi-dimensional data; possible uses include: • 2D Images, 3D turbulence data or multi-attribute sensor readings • Using wavelets in some example applications • Lossy compression, De-noising for images, Self-similarity analysis • Studying the phases of the wavelet filters (that delays the wavelet smoothes) and approximately computing the delay amount using DSP methods • Using this on Mammogram reconstruction • Possible uses of Bayesian? (not done)
DWTR / IDWTR wrappers • Assume D dimensions • Perform D sweeps, one across each dimension, making recursive calls for each D-1 dimensional slice • Top level recursive calls go D-1 levels deep before calling the 1 dimensional wavelet transformation functions • As a result 2^D-1 detail groups and a single smooth group is constructed for each level of transformation 7 detail groups smoothes smoothes 3 detail groups
Example Applications: Self-similarity Analysis • Calculate the means of the detail squares for each level and plot their log as a function of level • If the line is linear, then there is self-similarity • Brownian motion is self-similar, Random data (of course) is not
Mammogram Reconstruction • Assume all details are zero • Perform inverse wavelet transformation • Possible use of Bayesian Methods: • Model missing details using a Bayesian approach Original Image after wavelet interpolation after fixing delay problem
DSP Perspective: Problems Related with Non-zero Phase Filtering • Filtering in time domain is multiplication in frequency domain • Phase(Y(f)) = Phase(H(f))+Phase(X(f)) h[n] X[n] y[n]
Non-zero Phase Filtering • cos(2pf0t+f) = cos(2pf0(t+f/(2pf0)) • td = f/(2pf0) • td is constant if f is a linear function of frequency • Therefore, wavelet filters should be (approximately) linear phase filters • Symmetric filters have linear phase • Ex: {1, 1} (Haar), {1, 2, 1}
Least Asymmetric (LA) Wavelet Filters • Choose filter coefficients: s.t. min |f(f) – 2pfv| -L/2+1, if L =8,12,16,20 v = -L/2, if L =10, 18 -L/2+2, if L =14 • LA(8) and LA(12) works best.
The End! • Thanks!!!