DIGITAL IMAGE PROCESSING
Filtering in the
Frequency Domain
by
Dr. K. M. Bhurchandi
The DFT of One - Variable
• Obtaining the DFT from the Continuous Transform of a Sampled
Function
sampled, continuous,
band-limited function FT periodic function
extending from -∞ to ∞ extending from -∞ to ∞
• Equation of transform of sampled data in terms of sampled function is
given by:
2
• And we know
The DFT of One - Variable
• Although fn is a discrete function, its Fourier is continuous and infinitely periodic 3
with period 1/ ΔT.
• Sampling one period is the basis for the DFT.
The DFT of One - Variable
• To have M equally spaced samples of the transform taken over the period
μ = 0 to 1/ ΔT, consider following frequencies,
where, m = 0, 1, 2, ….., M-1
Substituting it in previous equation, we get DFT equation:
Conversely, IDFT can be obtained by:
where, n = 0, 1, 2, …, N-1
The DFT of One - Variable
• In a more intuitive approach, notations used will be
x & y - image coordinates variables
u & v – frequency variables
Thus, the above equations become:
where, u = 0, 1, 2, …., M-1
where, x = 0, 1, 2, …, M-1
5
• Both the forward & inverse discrete transforms are infinitely periodic with
period M.
Extension to Functions of Two Variables
• The 2-D Impulse and Its Sifting Property:
• The impulse of 2 continuous variables t & z is given by:
• &
• As in the 1D case, the 2D impulse exhibits the sifting property
under integration,
• More generally for an impulse located at coordinates (to, zo) 6
Extension to Functions of Two Variables
• For discrete variables x & y, the 2D discrete impulse is defined
as:
• Its sifting property is:
• For an impulse located at coordinates (xo, yo)
7
Extension to Functions of Two Variables
• Two-dimensional unit discrete impulse.
8
The 2-D Continuous Fourier Transform
Pair
• f(t, z) – continuous function of two variables, t & z.
• μ & γ – frequency variables define continuous frequency
domain.
9
• A 2-D function & its spectrum
10
2-D Sampling and the 2-D Sampling
Theorem
• 2-D impulse train is given by:
where, ΔT & ΔZ are separations between samples along t
& z axis.
11
2-D Sampling and the 2-D Sampling
Theorem
• Function f(t, z) is said to be band-limited if its Fourier
Transform is 0 outside a rectangle established by the intervals
[-μmax, μmax] and [-γmax, γmax].
• 2-Dimesional sampling theorem states that a continuous, band
limited function f(t, z) can be recovered with no error from a
set of its samples if the sampling intervals are
12
and
Aliasing in Images
• 2-D Fourier Transform of oversampled & under-sampled band
limited function
13
Aliasing in Images
• Two manifestations of aliasing in images:
• Spatial Aliasing & Temporal Aliasing
• Spatial Aliasing: It is due to under-sampling
• Temporal Aliasing: It is related to time intervals between
images in a sequence of images.
• E.g. wagon wheel effect (wheel appear to rotate backwards)
14
• Effects of Aliasing
15
The 2-D DFT and Its Inverse
• 2-D Discrete Fourier Transform Pairs:
where, f(x, y) – digital image of size M x N.
u – 0, 1, 2, ….., M-1
v – 0, 1, 2, ….., N-1
where, F(u, v) – transform of f(x, y). 16
x – 0, 1, 2, ….., M-1
y – 0, 1, 2, …., N-1
Properties
• of 2-D DFT
1) Relationships between Spatial & Frequency Intervals
• If a continuous function f(t, z) is sampled to form a digital
image, f(x, y) consisting of M x N samples in t& z resp.
• If ΔT&ΔZ denote the separations between samples, then,
separations between the corresponding discrete, frequency
domain variables are given by:
&
17
Properties of 2-D DFT
2) Translation and Rotation
• Fourier transform pair satisfies translation properties
Multiplying f(x,y) by exponential shown shifts the origin of DFT
to (u0,v0) and conversely, multiplying F(u,v) by negative of that
exponential shifts the origin of f(x,y) to (x0,y0).
18
Properties of 2-D DFT
2) Translation and Rotation
Using Polar coordinates
Results in transform pair
Rotating f(x,y) with rotates F(u,v) by same angle and viceversa. 19
Properties of 2-D DFT
3) Periodicity
2D Fourier transform and its inverse are infinitely periodic in u
and v directions
and
where k1 and k2 are integers
20
Properties of 2-D DFT
3) Periodicity
21
Problems on DFT
1)• Let f(x, y) = is image matrix. Then DFT(F) = ?
Sol:- y
f(x, y) = x =>
Generally, F(v) =
Thus, F1(0, v) =
22
• F1(0, v) =
F1(0, v) = + +
For, v = 0: F1(0, 0) = + +
=1x1 + 1x1 + 1x1 = 3
For, v = 1: F1(0, 1) = + +
= 1 x 1 + 1 x (-0.5000 - 0.8660i) + 1 x (-0.5000 + 0.8660i)
=0
For, v = 2: F1(0, 2) = + +
= 1 x 1 + 1 x (-0.5000 + 0.8660i) + 1 x (-0.5000 - 0.8660i)
=0
So, F1(0, v) = [3 0 0];
Similarly, F1(1, v) = [3 0 0]; & F1(2, v) = [3 0 0];
Thus, F1=
23
Now, applying transformation column-wise.
•Generally,
Thus,
+
For u=0: +
=3 x 1 +3 x 1+ 3 x 1
=9
For u=1: +
=3 x 1 +3 x (-0.5000-0.8660i)+ 3 x (-0.5000+0.8660i)
=0
For u=2: +
=3 x 1 +3 x (-0.5000+0.8660i)+ 3 x (-0.5000-0.8660i)
=0
So, F(u,0)=
Similarly, F(u,1)=; & F(u,2)=
Thus, F=
24
Problems on DFT
2)• Let
f(x, y) = is image matrix. Then DFT(F) = ?
Sol:- y
f(x, y) = x =>
Generally, F(v) =
Thus, F1(0, v) =
25
• F1(0, v) =
F1(0, v) = + +
For, v = 0: F1(0, 0) = + +
=1x1 + 2x1 + 3x1 = 6
For, v = 1: F1(0, 1) = + +
= 1 x 1 + 2 x (-0.5000 - 0.8660i) + 3 x (-0.5000 + 0.8660i)
= -1.5000 + 0.8660i
For, v = 2: F1(0, 2) = + +
= 1 x 1 + 2 x (-0.5000 + 0.8660i) + 3 x (-0.5000 - 0.8660i)
= -1.5000 - 0.8660i
So, F1(0, v) = [6 -1.5000 + 0.8660i -1.5000 - 0.8660i];
Similarly, F1(1, v) = [15 -1.5000 + 0.8660i -1.5000 - 0.8660i];
&F1(2, v) = [24 -1.5000 + 0.8660i -1.5000 - 0.8660i]; 26
Thus, F1=
•
Generally, F(u) =
Thus, F(u, 0) =
F(u, 0) = F1 + +F1
For, u = 0: F(0, 0) = F1 +F1 +F1
= 6 x 1 + 15 x 1 + 24 x 1 = 45
For, u = 1: F(1, 0) = F1 +F1 +F1
= 6x 1 + 15 x (-0.5000 - 0.8660i) + 24 x (-0.5000 + 0.8660i)
= -13.5000 + 7.7940i
For, u = 2: F(2, 0) = F1 +F1 +F1
= 6 x 1 + 15 x (-0.5000 + 0.8660i) + 24 x (-0.5000 - 0.8660i)
= -13.5000 - 7.7940i
So, F(u,0)=
Similarly, F(u,1) =
& F(u,2) =
Thus, F=
27
3)• Let
f(x, y) = is image matrix. Then DFT(F) = ?
Sol:- y
f(x, y) = x
Generally, F(v) =
Thus, F1(-1, v) =
28
• wise calculation
Row
F1(-1, v) = + +
F1(0, v) = + +
F1(1, v) = + +
&
Column wise calulation
F(u, -1) = F1 + +F1
F(u, 0) = F1 + +F1
F(u, 1) = F1 + +F1
Then,
F=
F=
29
Symmetry Properties of 2D-DFT
30
Summary of DFT definitions
31
Summary of DFT definitions
32
Summary of DFT Pairs
33
Summary of DFT Pairs
34
DFT of an image
35
Spectrum is insensitive to translation. But
rotates by the same angle as the rotated
image
36
Translation affects the phase
37
Reconstruction using either
spectrum or phase
38
Frequency Domain Filtering
Fundamentals
• Given a digital image f(x,y) of size MxN, the basic filtering
equation is
Input image Frequency
Spectrum 39
Result of Filtering an image
More Results on Filtering
High Pass Filter
Low Pass Filter
40
Top row: Frequency domain Filters, Bottom Row: Corresponding Filtered Images
Gaussian Low pass Filtering
Gaussian Low
Gaussian Low
Pass Filter
Pass Filter with
without padding
padding
41
Summary of steps for filtering in Frequency
domain
1) Given an input image f(x, y) of size M x N, obtain the padding parameters P & Q
as 2M & 2N respectively.
2) Form a padded image, fp(x, y), of size P x Q be appending the necessary number
of zeros to f(x, y).
3) Multiply fp(x, y) by (-1)x+y to center its transform.
4) Compute the DFT, F(u, v), of the image from step 3.
5) Generate a real, symmetric filter function, H(u, v), of size P x Q with center at
coordinates (P/2, Q/2). Form the product G(u, v) = H(u, v)F(u, v) using array
multiplication; that is, G(i, k) = H(i, k)F(i, k).
6) Obtain the processed image: gp(x, y) = {real[Ṫ-1[G(u, v)]]}(-1)x+y where the real
part is selected in order to ignore parasitic complex components resulting from
computational inaccuracies, and the subscript p indicates that we are dealing
with padded arrays. 42
7) Obtain the final processed result, g(x, y), by extracting the M x N region from the
top, left quadrant of gp(x, y).
• Steps for Filtering in the Frequency Domain
43
Image smoothing using Frequency domain
filters
• Ideal
Low-pass Filters (ILPF):
• A 2-D low-pass filter that passes without attenuation all
frequencies within a circle of radius D0 from the origin and
“cut off” all frequencies outside this circle is called an ideal
low-pass filter (ILPF).
H
Where, D0 - a positive constant
D(u, v) – distance between (u, v) in the frequency domain and
the center of the frequency rectangle; i.e.
D(u, v) = [(u – P/2)2 + (v – Q/2)2]1/2
44
Where P & Q are 2M & 2N respectively.
Image smoothing using Frequency domain
filters
Perspective plot
of an Ideal LPF
Transfer function
45
Filter displayed as an image Filter radial cross section
• Ideal Low Pass Filtering
46
Results of
Ideal LPF
Original Result of filtering
Image with ideal low pass
filter of radius 10
Result of filtering Result of filtering
with ideal low pass with ideal low pass
filter of radius 30 filter of radius 60
Result of filtering
with ideal low pass Result of filtering
with ideal low pass 47
filter of radius 160
filter of radius 460
The blurring and “ringing” properties
of Ideal LPFs
Intensity profile of a horizontal
line passing through the
center of the image.
48
Representation in the spatial
domain of an ILPF of radius 5
and size 1000x1000.
Image smoothing using Frequency domain
filters
• Butterworth
Low-pass Filters (BLPF):
• The transfer function of a Butterworth low-pass filter (BLPF) of
order n, and with cutoff frequency at a distance D0 from the
origin, is defined as,
H
where, D(u, v) = [(u – P/2)2 + (v – Q/2)2]1/2
49
Image smoothing using Frequency domain
filters
Perspective
from order 1 - 4 plot of BLPF transfer
function
50
Filter displayed as an image Filter Radial cross section of
orders 1 through 4
Image smoothing using Frequency domain
filters
• Spatial representation of BLPF of order 1, 2, 5 & 20 and
corresponding intensity profiles through the center of the
filter.
• Ringing increases as a function of filter order.
51
Results of Butterworth LPF
Result of filtering
Original with BLPF with
Image cutoff radius 10
Result of Result of filtering
filtering with with BLPF with
BLPF with cutoff cutoff radius 60
radius 30
Result of Result of filtering
filtering with with BLPF with 52
BLPF with cutoff cutoff radius 460
radius 160
Image smoothing using Frequency domain
filters
• Gaussian
Low-pass Filters (GLPF):
• Gaussian low-pass filters(GLPF) in two dimensions is given by
where, σ – is measure of spread about the center.
• Let σ = D0, thus the filter can be expressed as:
53
where D0 is the cutoff frequency
Image smoothing using Frequency domain
filters
Perspective
Plot of Gaussian LPF
transfer function
54
Filter displayed as Filter radial cross section
an image for various values of D0
Results of Gaussian Low pass Filters
Input Image Result of filtering
with GLPF with cut
off radius 10
Result of
filtering with Result of filtering
GLPF with cut with GLPF with cut
off radius 30 off radius 60
Result of
Result of filtering 55
filtering with
with GLPF with cut
GLPF with cut
off radius 460
off radius 160
Low Pass Filters
56
Applications of LPF
• Character Recognition
57
Result of filtering with GLPF
Sample text of low resolution
Note: broken character
Note: broken characters
segments are joined
Contd..
• Printing and publishing industry: “cosmetic” processing
58
Image sharpening using Frequency
domain filter
• A
high-pass filter is obtained from a given low-pass filter using
the equation
HHP(u, v) = 1 – HLP(u, v)
Where, HLP(u, v) – transfer function of the LPF
• Ideal High-pass Filter (IHPF)
• A 2-D high-pass filter is given by
59
H
Image sharpening using Frequency
domain filter
• Butterworth
High-pass Filter (BHPF)
• A 2-D Butterworth high-pass filter (BHPF) of order n and cutoff
frequency D0 is defined as
H
• Gaussian High-pass Filter (GHPF)
• The transfer function of the Gaussian high-pass filter (GHPF)
with cutoff frequency locus of at a distance D0 from the center
of the frequency rectangle is given by
60
Image sharpening using Frequency
domain filter
IHPF
BHPF
61
GHPF
Spatial Representation of High Pass Filter
62
Results of Ideal HPF
63
Results of BHPF of order 2
• Results are much smoother than ideal HPF
64
Results of GHPF
𝐷 0= 30 𝐷 0= 60 𝐷 0= 160
65
Laplacian in the frequency domain
• The
Laplacian can be implemented in the frequency domain
using the filter
Or with respect to the center of the frequency rectangle, using
the filter
]
=
Where, D(u, v) is the distance function. 66
Laplacian in the frequency domain
• Laplacian image is obtained as:
The
Enhancement is achieved using the equation
• Here, c = -1, because, H(u, v) is negative
• In frequency domain the above equation is written as:
67
Homomorphic Filtering
• According to Illumination-Reflectance model:
• An image f(x, y) can be expressed as the product of its illumination,
i(x, y), and reflectance, r(x, y), components:
f(x, y) = i(x, y)r(x, y) (1)
• Since the Fourier Transform of a product is not the product of the
transforms:
• So, F[f(x, y)] ≠ F[i(x, y)] F[r(x, y)]
• However, if,
z(x, y) = ln f(x, y)
= ln i(x, y) + ln r(x, y) (2)
Then, 68
F{z(x, y)} = F{ln f(x, y)}
= F{ln i(x, y)} + F{ln r(x, y)} (3)
Homomorphic Filtering
or
Z(u, v) = Fi(u, v) + Fr(u, v)
Where, Fi(u, v) and Fr(u, v) are the Fourier transforms of ln i(x, y) & ln
r(x, y) respectively.
We can filter Z(u, v) using a filter H(u, v)
S(u, v) = H(u, v) Fi(u, v) + H(u, v) Fr(u, v) (4)
The filtered image in spatial domain is
s(x, y) = T-1{S(u, v)}
= T-1{H(u, v) Fi(u, v)} + T-1{H(u, v) Fr(u, v)} (5)
= i’(x, y) + r’(x, y)
69
where, i’(x, y) = T-1{H(u, v) Fi(u, v)}
r’(x, y) = T-1{H(u, v) Fr(u, v)}
Homomorphic Filtering
• In order to get the original image, we reverse the process,
g(x, y) = es(x, y)
= ei’(x, y)er’(x, y)
= i0(x, y)r0(x, y) (6)
where, i0(x, y) = ei’(x, y) & r0(x, y) = er’(x, y) are the illumination and
reflectance components of the output (processed) image.
70
Summary of steps in homomorphic
filtering
• The illumination component of an image is characterized by
slow spatial variations, while the reflectance components tend
to vary abruptly.
• The low frequencies of DFT of Logarithm of an image are
71
associated with illumination and high frequencies with
reflectance.
Homomorphic Filtering
• Radial cross section of such a homomorphic filter H(u, v) that
affects low and high frequency components differently. The vertical 72
axis is at the center of the frequency rectangle and D(u,v) is the
distance from the center.
Homomorphic Filtering
• If
parameters γL&γH are chosen such that γL< 1 &γH> 1, the filter
attenuates low frequencies and amplifies high frequencies.
• It results in simultaneous dynamic range compression and
contrast enhancement. (High-pass filter)
• Slightly modified Gaussian high-pass filter yields the function
]+
where, constant ‘c’ controls the sharpness of the
slope of the function as it transitions between
73
γL&γH