Digital Image Processing

Image Enhancement in Frequency

Domain
(Fundamentals)

02/02/25 BCSE3045 Digital Image Processing 1

 Filtering in the Frequency Domain
Filter: A device or material for suppressing or
minimizing waves or oscillations of certain frequencies.

Frequency: The number of times that a periodic

function repeats the same sequence of values during a
unit variation of the independent variable.

Webster’s New Collegiate Dictionary

02/02/25 BCSE3045 Digital Image Processing 2

 Jean Baptiste Joseph Fourier
Fourier was born in Auxerre,
France in 1768.
– Most famous for his work “La
Théorie Analitique de la Chaleur”
published in 1822.
– Translated into English in 1878:
“The Analytic Theory of Heat”.
Nobody paid much attention when the work was
first published.
One of the most important mathematical theories
in modern engineering.
02/02/25 BCSE3045 Digital Image Processing 3
 The Big Idea
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

Any function that periodically repeats itself can

be expressed as a sum of sines and cosines of
different frequencies each multiplied by a
different coefficient – a Fourier series
02/02/25 BCSE3045 Digital Image Processing 4
 1D continuous signals
• It may be
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

considered both as
continuous and
discrete.
• Useful for the
representation of , x  x0
discrete signals  ( x  x0 ) 
 0 otherwise
through sampling

of continuous
signals.
f ( x) ( x  x )dx  f ( x )

0 0

02/02/25 BCSE3045 Digital Image Processing 5

 1D continuous signals (cont.)
Impulse train function
Images taken from Gonzalez & Woods, Digital Image Processing (2002)


S T (t )    (t  nT )
n  

 
x[n]  x(t ) S T (t )   x(t ) (t  nT )   x(nT ) (t  nT )
n   n  

02/02/25 BCSE3045 Digital Image Processing 6

 1D continuous signals (cont.)
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

x[n]  x(t ) S T (t )


  x(t ) (t  nT )
n  


  x(nT ) (t  nT )
n  

02/02/25 BCSE3045 Digital Image Processing 7

 1D continuous signals (cont.)
• The Fourier series expansion of a periodic
signal f (t).
 2
j nt
f (t )   cn e T

n  

T /2 2
1  j nt
cn   f (t )e T
dt
T  T /2
02/02/25 BCSE3045 Digital Image Processing 8
 1D continuous signals (cont.)
• The Fourier transform of a continuous
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

signal f (t).

F (  )   f (t )e  j 2 t
dt


f (t )  F (  )e j 2 t
d

• Attention: the variable is the frequency (Hz) and
not the radial frequency (Ω=2πμ) as in the Signals
and Systems course.
02/02/25 BCSE3045 Digital Image Processing 9
 1D continuous signals (cont.)
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

sin(W )
f (t ) PW / 2 (t )  F (  ) W
(W )
02/02/25 BCSE3045 Digital Image Processing 10
 1D continuous signals (cont.)
• Intermediate result
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

− The Fourier transform of the impulse train.

 
1  n 

n  
 (t  nT ) 
T
  
n   

T 

• It is also an impulse train in the frequency

domain.
• Impulses are equally spaced every 1/ΔΤ.

02/02/25 BCSE3045 Digital Image Processing 11

 1D continuous signals (cont.)
Sampling
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

x[n]  x(t ) S T (t )


  x(t ) (t  nT )
n  


  x(nT ) (t  nT )
n  

02/02/25 BCSE3045 Digital Image Processing 12

 1D continuous signals (cont.)
• Sampling
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

− The spectrum of the discrete signal consists of

repetitions of the spectrum of the continuous
signal every 1/ΔΤ.
− The Nyquist criterion should be satisfied.

f (t )  F (  )

f (nT )  f [n]  F (  )  1  n 
T

n  
F


T 
02/02/25 BCSE3045 Digital Image Processing 13
 The Discrete Fourier Transform
• The Fourier transform of a sampled (discrete)
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

signal is a continuous function of the frequency.


 1  n 
F ( ) 
T

n  
F
 T


• For a N-length discrete signal, taking N samples of
its Fourier transform at frequencies:
k
k  , k 0,1, .., N  1
N T
provides the discrete Fourier transform (DFT) of
the signal.
02/02/25 BCSE3045 Digital Image Processing 14
 The Discrete Fourier Transform (cont.)
• DFT pair of signal f [n] of length N.
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

N1 2 nk
 j
F [k ]  f [n]e N
, 0 k  N  1
n 0
N1 2 nk
1 j
f [ n] 
N
 F [k ]e
n 0
N
, 0 n  N  1

02/02/25 BCSE3045 Digital Image Processing 15

 The Discrete Fourier Transform (cont.)
• Property
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

– The DFT of a N-length f [n] signal is periodic

with period N.
F [ k  N ] F [ k ]
– This is due to the periodicity of the complex
exponential:
2 n 2 n
  j   j
wN 
n n
 e N
  wN e
N

 
02/02/25 BCSE3045 Digital Image Processing 16
 The Discrete Fourier Transform (cont.)
• Property: sum of complex exponentials
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

1 N
1, k rN , r  
 kn
w 
N
N n 0 0, otherwise

The proof is left as an exercise.

2
 j
wN e N

02/02/25 BCSE3045 Digital Image Processing 17

 The Discrete Fourier Transform (cont.)
• DFT pair of signal f [n] of length N may be
Images taken from Gonzalez & Woods, Digital Image Processing (2002)

expressed in matrix-vector form.

N1
F [k ]  f [n]w , nk
N 0 k  N  1
n 0
N1
1
f [ n] 
N
 F [ k
n 0
]w  nk
N , 0 n  N  1

2
 j
wN e N

02/02/25 BCSE3045 Digital Image Processing 18

The Discrete Fourier Transform (cont.)
F = Af
 w0 0 w  0 1
w  0 2
 w 
0 N1

 N N N N

 1 0 1 N1 
 wN  w  1 1
w  w 
1 2
  
A N N N
      
 
 w N  1 0 w  w 
N1 1 N1 2
 wN  
N1 N1 
 N N N

f =  f [0], f [1],..., f [ N  1] , F =  F [0], F [1],..., F [ N  1]

T T

02/02/25 BCSE3045 Digital Image Processing 19

The Discrete Fourier Transform (cont.)
Example for N=4

1 1 1 1 
1  j  1 j 
A  
1  1 1  1
 
1 j  1  j 

02/02/25 BCSE3045 Digital Image Processing 20

The Discrete Fourier Transform (cont.)
The inverse DFT is then expressed by:
f = A -1F
* T
 
 N w  w  w  
00 0 1 0 2 0 N1
w N N  N 
 1 0 1 N1 

1  w  w 1 1
w  wN  
1 2
1  
A 1 
N
 A 
* T
  N
N  
N N
 
     

  N1 0 
  wN   w  w  wNN  1  
N1 1 N1 2 N  1
N N  
 

This is derived by the complex exponential sum

property.
02/02/25 BCSE3045 Digital Image Processing 21