Lustig eecs uc berkeley m lustig eecs uc berkeley 17

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: function M. Lustig, EECS UC Berkeley M. Lustig, EECS UC Berkeley 17 18 Haar Scaling function 8 < 1 0t< 1 2 1 1 t<1 ( t) = 2 : 0 otherwise ( t) = ⇢ Back to Discrete • Early 80’s, theoretical work by Morlett, Grossman and Meyer (math, geophysics) • Late 80’s link to DSP by Daubechies and Mallat. 1 0t<1 0 otherwise • From CWT to DWT not so trivial! • Must take care to maintain properties M. Lustig, EECS UC Berkeley M. Lustig, EECS UC Berkeley 19 20 Discrete Wavelet Transform d s ,u = a s ,u = N1 X n=0 N1 X Discrete Wavelet Transform x [ n] x [ n] s ,u [ n] d s ,u = s ,u [ n] a s ,u = n=0 d00 N1 X n=0 N1 X x [ n] s ,u [ n] x [ n] s ,u [ n] n=0 d01 d02 d00 d03 ! d01 d02 d03 ! d10 d11 d10 a10 d20 a20 finest scale d11 a11 stop here: n n M. Lustig, EECS UC Berkeley M. Lustig, EECS UC Berkeley 21 22 Example: Discrete Haar Wavelet Haar for n=2 Discrete Orthogonal Haar Wavelet Haar for n=8 1 p 2 scaling Φ20 1 p 2 scaling function Ψ00 1 p 8 1 p 2 0 ,0 0,0 1 p 2 Ψ20 Ψ01 1 p 8 1 p 2 mother wavelet Ψ10 approximation detail d00 Ψ11 ! Ψ02 1 2 1 p 2 1 2 Ψ03 a00 Equivalent to DFT2! ! t M. Lustig, EECS UC Berkeley 23 t M. Lustig, EECS UC Berkeley 24 Fast DWT with Filter Banks (more Later!) h0[n] x [ n] Fast DWT with Filter Banks h0[n] h1[n] h0[n] h1[n] h1[n] not quite... too many d0n? coefficien...
View Full Document

Ask a homework question - tutors are online