Beruflich Dokumente
Kultur Dokumente
= 3
+5
+ 10
+ 23
+ ...
2 5 0
1 8 3
3 7 5
0 0 0
0 0 0
1 0 0
0 0 0
0 0
+1
0 0
0 0 0
1 0 0
0 0 0
0 0 0
+3
0 0
+5
1 0
+ ...
f ( x ) = F ( ) e
i 2x
Standard Basis:
[ a1 a2 a3 a4 ] =
a1 [ 1 0 0 0 ] + a2 [ 0 1 0 0 ] + a3 [ 0 0 1 0 ] + a4 [ 0 0 0 1 ]
Hadamard Transform:
[ 2 1 0 1] =
= [ 1 1/2 -1/2 0 ] Hadamard
= 1 [ 1 1 1 1 ] + 1/2 [ 1 1 -1 -1 ] - 1/2 [ -1 1 1 -1 ] + 0 [ -1 1 -1 1 ]
1. Basis Functions. 2. Method for finding the image given the transform coefficients. 3. Method for finding the transform coefficients given the image.
standard
T0 = [ 1 1 1 1 ] T1 = [ 1 1 -1 -1 ] T2 = [ -1 1 1 -1 ] T3 = [ -1 1 -1 1 ]
New Coefficients: a0 = <X,T0 > =< [ 2 1 0 1 ] , [ 1 1 1 1 ] > /4 = 1 a1 = <X,T1 > =< [ 2 1 0 1 ] , [ 1 1 -1 -1 ] > /4 = 1/2 a2 = <X,T2 > =< [ 2 1 0 1 ] , [ -1 1 1 -1 ] > /4 = -1/2 a3 = <X,T3 > =< [ 2 1 0 1 ] , [ -1 1 -1 1 ] > /4 = 0
Signal:
X = [ 1 1/2 -1/2 0 ]
new
Grayscale Image
Y Frequency
Fourier Image
X Coordinate
X Frequency
= a11
[ ]
1 0 0 0
+ a12
[ ] [ ] [ ]
0 1 0 0 0 0 + a21 + a22 0 0 1 0 0 1
Hadamard Transform:
[ ]
1 0
= 1
[ ] [ ] [ ] [ ] [ ]
1 1 1 -1 -1 -1 -1 1 + 1/2 - 1/2 + 0 1 1 1 -1 1 1 1 -1 1 1/2 -1/2 0
Hadamard
1. Basis Functions. 2. Method for finding the image given the transform coefficients. 3. Method for finding the transform coefficients given the image.
Standard Basis:
2 1
1 0
] [ ] [ ] [ ] [ ]
= 2 1 0 0 0 + 1 0 1 + 1 0 0 0 0 + 0 1 0 0 0 0 1 1 0
2 1
[ ] [ ]
1 0 0 0 0 1 0 0 0 0 0 1
coefficients
[ ] [ ]
0 0 1 0
Standard
1/2 0
-1/2
] [ ]
= 1
1 1 1 -1 -1 -1 -1 1 + 1/2 - 1/2 + 0 1 1 1 -1 1 1 1 -1
[ ] [ ] [ ] [ ] [ ]
1 1 1 1 1 -1 1 -1
1 1/2 -1/2 0
Hadamard
coefficients
[ ] [ ]
-1 -1 1 1 -1 1 1 -1
Hadamard
Basis Elements
size = 8x8
Black = +1 White = -1
f ( x) =
2) The dot product:
a
i
B i (x) di
<f(x),Bi(x)> =
f(x) B i (x) dx
Fourier Transform
Basis Functions are sines and cosines
sin(x)
cos(2x)
sin(4x)
a sin(2x)
2a sin(2x)
-a sin(2x)
3 sin(x)
+ 1 sin(3x)
B A+B
+ 0.8 sin(5x)
C A+B+C D A+B+C+D
+ 0.4 sin(7x)
f (x) =
F ( ) e i 2 x d
F ( ) = f(x)e
x
i2 x
dx
~ F { f ( x )} = F ( )
F ( ) = R e
R tells us how much of frequency is needed. tells us the shift of the Sine wave with frequency .
Alternatively:
F ( ) = a + ib
a tells us how much of cos with frequency is needed. b tells us how much of sin with frequency is needed.
R - is the amplitude of F(). - is the phase of F(). |R|2=F* () F() - is the power spectrum of F() . If a signal f(x) has a lot of fine details F() will be high for high . If the signal f(x) is "smooth" F() will be low for high .
Fourier Transform
Original Problem
Difficult solution
Examples:
The Delta Function: Let
f ( x) = (x )
F ( ) =
f(x)
i 2 x ( x ) e =1
Fourier
f ( x ) =1
i 2 x ( ) F = e = ()
f(x)
Fourier
Real
Imag
0
f ( x) = cos(2 0 x )
1 i 20 x i 20 x i 2x F ( )= e +e e dx = 2 1 = [ ( 0 )+ ( + 0 )] 2
f(x)
Fourier
x Real
-0
-0
Imag
F ( ) = e
0.5
sin ( ) dx = = sinc ( )
f(x)
x
-0.5 Fourier 0.5
Proof:
f(x) = rect1/2(x) =
1/2
1 0
-1/2
1/ 2
[ 2iu
sinc(u)
f ( x) = e
x 2
F ( )= e
f(x)
x
Fourier
f(x)
ck(x)
x
k
Fourier
C1/k()
1/k
Fourier Transform - 2D
Given a continuous real function f(x,y), its Fourier transform F(u,v) is defined as:
~ F{f (x,y)}= F(u,v) = f (x,y)e 2i(ux+vy)dxdy
Phase =
(u,v) = tg-1(b(u,v)/a(u,v))
Spectrum (Amplitude) = |F(u,v)| = a2(u,v) + b2(u,v) Power Spectrum = |F(u,v)|2 = a2(u,v) + b2(u,v)
1 u2 + v2
1 v 1/u
u=0 v=0
direction of waves
direction of waves
u=-2, v=2
u=-1, v=2
u=0, v=2
u=1, v=2
u=2, v=2
u=-2, v=1
u=-1, v=1
u=0, v=1
u=1, v=1
u=2, v=1
U
u=-2, v=0 u=-1, v=0 u=0, v=0 u=1, v=0 u=2, v=0
u=-2, v=-1
u=-1, v=-1
u=0, v=-1
u=1, v=-1
u=2, v=-1
u=-2, v=-2
u=-1, v=-2
u=0, v=-2
u=1, v=-2
u=2, v=-2
2D Fourier Transform
f(x,y)
1 1/2 1/2
f(x,y) = rect(x,y) =
|F(u,v)|
F (u ) =
f ( x, y )e 2i ( ux + vy ) dxdy =
1 / 2 1/ 2
1 / 2 1 / 2
2i ( ux + vy ) e dxdy
1/2
1 / 2
e 2 iux dx
1/2
1 / 2
e 2 ivy dy