Lect07 Image Restoration

Digital Image Processing
Lecture # 07
Image Restoration
Outline
Noise in the context of image processing
Noise Modelling
Noise removal techniques are typically used in image

processing?
Deblurring techniques are typically used in image

processing?
Lecture # 7
Image Restoration
Image restoration: recover an image that has been

degraded by using a prior knowledge of the degradation
phenomenon.
Model the degradation and applying the inverse process in

order to recover the original image.
Lecture # 7
A Model of Image Degradation/Restoration

Process
Lecture # 7

Process
If H is a linear,
position-invariant process, then
the degraded image is given in the spatial domain by

g ( x , y ) h ( x , y ) f ( x , y ) ( x, y )
Lecture # 7

Process
The model of the degraded image is given in

the frequency domain by
G (u , v) H (u , v) F (u, v) N (u, v)
Lecture # 7
Noise Sources
The principal sources of noise in digital images arise during

image acquisition and/or transmission
Image acquisition
e.g., light levels, sensor temperature, etc.
Transmission
e.g., lightning or other atmospheric disturbance in wireless
network
Lecture # 7
Noise Models (1)
Gaussian noise
Electronic circuit noise, sensor noise due to poor illumination and/or high
temperature
Rayleigh noise
Range imaging
Erlang (gamma) noise: Laser imaging
Exponential noise:
Uniform noise: Least descriptive;
Impulse noise: quick transients,
Laser imaging
generators
Basis for numerous random number
such as faulty switching
Lecture # 7
Lecture # 7
10
Gaussian Noise (1)

The PDF of Gaussian random variable, z, is given by
1
( z z )2 /2 2
p(z )
e
2
where, z represents intensity
z is the mean (average) value of z
is the standard deviation
Lecture # 7
17
Gaussian Noise (2)

The PDF of Gaussian random variable, z, is given by
1
( z z )2 /2 2
p(z )
e
2
70% of its values will be in the range
( ), ( )
95% of its values will be in the range
( 2 ), ( 2 )
Lecture # 7
18
Rayleigh Noise
The PDF of Rayleigh noise is given by
2
( z a )2 / b
for z a
( z a )e
p( z ) b
0
for z a
The mean and variance of this density are given by
z a b / 4
b(4 )
2

4
Lecture # 7
19
Erlang (Gamma) Noise

The PDF of Erlang noise is given by
ab z b1 az
e
p( z ) (b 1)!
0
for z 0
for z a

z b/a
2 b / a2
Lecture # 7
20
Exponential Noise
The PDF of exponential noise is given by
ae az
p( z )
0
for z 0
for z a

z 1/ a
2 1/ a 2
Lecture # 7
21
Uniform Noise
The PDF of uniform noise is given by
1
p( z ) b a
0
for a z b
otherwise

z (a b) / 2
2 (b a)2 /12
Lecture # 7
22
Impulse (Salt-and-Pepper) Noise

The PDF of (bipolar) impulse noise is given by
Pa
p( z ) Pb
0
for z a
for z b
otherwise
if b a, gray-level b will appear as a light dot,

while level a will appear like a dark dot.
If either Pa or Pb is zero, the impulse noise is called
unipolar
Lecture # 7
23
Examples of Noise: Original Image
Lecture # 7
24
Examples of Noise: Noisy Images(1)
Lecture # 7
25
Examples of Noise: Noisy Images(2)
Lecture # 7
26
Effects of
noise on
images and
histograms
Gaussian
Exponential
Impulse
(salt-andpepper)
Lecture # 7
27
Effects of
noise on
images and
histograms
Rayleigh
Gamma
(Erlang)
Uniform
Lecture # 7
28
Periodic Noise
Periodic noise in an image arises typically from electrical or

electromechanical interference during image acquisition.
It is a type of spatially dependent noise
Periodic noise can be reduced significantly via frequency

domain filtering
Lecture # 7
29
An Example of Periodic Noise
Lecture # 7
30
Noise estimation
Lecture # 7
31
Estimation of Noise Parameters (1)

The shape of the histogram identifies the closest PDF match
Lecture # 7
32
Estimation of Noise Parameters (2)

Consider a subimage denoted by S , and let ps ( zi ), i 0, 1, ..., L -1,
denote the probability estimates of the intensities of the pixels in S .
The mean and variance of the pixels in S:
L 1
z zi ps ( zi )
i 0
L 1
and
2 ( zi z )2 ps ( zi )
i 0
Lecture # 7
33
Restoration in the Presence of Noise Only
Spatial Filtering
Noise model without degradation
g ( x, y ) f ( x, y ) ( x, y )
and
G (u, v) F (u, v) N (u, v)
Lecture # 7
34
Spatial Filtering: Mean Filters (1)

Let S xy represent the set of coordinates in a rectangle
subimage window of size m n, centered at ( x, y ).
Arithmetic mean filter

1
f ( x, y )
g (s, t )
mn ( s ,t )Sxy
Lecture # 7
35

Geometric mean filter
f ( x, y ) g ( s , t )
( s ,t )S xy
1
mn
Generally, a geometric mean filter achieves smoothing

comparable to the arithmetic mean filter, but it tends to
lose less image detail in the process
Lecture # 7
36

Harmonic mean filter
f ( x, y )
mn
( s ,t )S xy
1
g ( s, t )
It works well for salt noise, but fails for pepper noise.
It does well also with other types of noise like Gaussian
noise.
Lecture # 7
37

Contraharmonic mean filter
f ( x, y )
g ( s, t )Q 1
g ( s, t )Q
( s ,t )S xy
( s ,t )S xy
Q is the order of the filter.

It is well suited for reducing the effects of salt-andpepper noise. Q>0 for pepper noise and Q<0 for salt
noise.
Lecture # 7
38
Spatial Filtering: Example (1)
39
Lecture # 7
39
11/5/2015
40
Lecture # 7
40
Lecture # 7
41
Noise reduction Mean filters
FIGURE: (a) Original

image; (b) image
with Gaussian noise;
(c) result of 3 3
arithmetic mean
filtering; (d) result of
5 5 arithmetic
mean filtering; (e)
result of 3 3
geometric mean
filtering; (f) result of
3 3 harmonic mean
filtering
Lecture # 7
42
Noise reduction Mean filters
FIGURE: (a) Image

with salt and pepper
noise; (b) result of 3
3 arithmetic mean
filtering; (c) result of
3 3 geometric
mean filtering; (d)
result of 3 3
harmonic mean
filtering; (e) result of
3 3 contraharmonic
mean filtering with R
= 0.5; (f) result of 3
3 contraharmonic
mean filtering with R
= 0.5.
Lecture # 7
43
Spatial Filtering: Order-Statistic Filters (1)

Median filter
f ( x, y ) median g (s, t )
( s ,t )S xy
Max filter
f ( x, y ) max g (s, t )
( s ,t )S xy
Min filter
f ( x, y ) min g (s, t )
( s ,t )S xy
Midpoint filter
1
f ( x, y ) max g ( s, t ) min g ( s, t )
( s ,t )S xy
2 ( s ,t )S xy
Alpha-trimmed mean filter

1
f ( x, y)
gr ( s, t )
mn d ( s ,t )Sxy
We delete the d / 2 lowest and the d / 2 highest intensity values of
g ( s, t ) in the neighborhood S xy . Let g r ( s, t ) represent the remaining
mn - d pixels.
11/5/2015
47
Spatial Filtering: Adaptive Filters (1)

Adaptive filters
The behavior changes based on statistical characteristics

of the image inside the filter region defined by the mn
rectangular window.
The performance is superior to that of the filters
discussed
Adaptive Filters:
Adaptive, Local Noise Reduction Filters (1)
S xy : local region
The response of the filter at the center point (x,y) of S xy
is based on four quantities:
(a) g ( x, y ), the value of the noisy image at (x, y );
(b) 2 , the variance of the noise corrupting f ( x, y )
to form g ( x, y );
(c) mL , the local mean of the pixels in S xy ;
(d) L2 , the local variance of the pixels in S xy .
Adaptive Filters:
The behavior of the filter:
(a) if 2 is zero, the filter should return simply the value
of g ( x, y ).
(b) if the local variance is high relative to 2 , the filter
should return a value close to g ( x, y);
(c) if the two variances are equal, the filter returns the
arithmetic mean value of the pixels in S xy .
Adaptive Filters:
An adaptive expression for obtaining f ( x, y)

based on the assumptions:
f ( x , y ) g ( x , y ) 2 g ( x , y ) mL
L
2
Adaptive Filters:
Adaptive Median Filters (1)
The notation:
zmin minimum intensity value in S xy
zmax maximum intensity value in S xy
zmed median intensity value in S xy
z xy intensity value at coordinates ( x, y )
Smax maximum allowed size of S xy
Adaptive Filters:
The adaptive median-filtering works in two stages:
Stage A:
A1 = zmed zmin ; A2 = zmed zmax
if A1>0 and A2<0, go to stage B
Else increase the window size
if window size Smax , repeat stage A; Else output zmed
Stage B:
B1 = z xy zmin ;
B2 = z xy zmax
if B1>0 and B2<0, output z xy ; Else output zmed
Adaptive Filters:
The adaptive median-filtering works in two stages:
Stage A:
A1 = zmed zmin ; A2 = zmed zmax
if A1>0 and A2<0, go to stage B
Else increase the window size
if window size Smax , repeat stage A; Else output zmed
Stage B:
B1 = z xy zmin ;
B2 = z xy zmax
if B1>0 and B2<0, output z xy ; Else output zmed
Example:
Adaptive Median Filters
Periodic Noise Reduction by Frequency
Domain Filtering
The basic idea
Periodic noise appears as concentrated bursts of energy

in the Fourier transform, at locations corresponding to
the frequencies of the periodic interference
Approach
A selective filter is used to isolate the noise
Lecture # 7
59
Perspective Plots of Bandreject Filters
Lecture # 7
60
Noise reduction frequency

domain
Bandreject
filter
attenuates frequency components within a

certain range (the stopband of the filter) while
leaving all other frequency components
untouched (or amplifying them by a certain
gain).
Ideal bandreject filter

Lecture # 7
61
Noise reduction frequency

domain
Bandreject
filter
Butterworth bandreject filter
Gaussian bandreject filter
Lecture # 7
62
A Butterworth bandreject
filter of order 4, with the
appropriate radius and
width to enclose
completely the noise
impulses
Lecture # 7
63
Linear, Position-Invariant Degradations

H is linear
H af1 ( x, y ) bf 2 ( x, y ) aH f1 ( x, y ) bH f 2 ( x, y )
f1 and f 2 are any two input images.
An operator having the input-output relationship
g ( x, y ) H f ( x, y ) is said to be position invariant
if
H f ( x , y ) g ( x , y )
for any f ( x, y ) and any and .
Lecture # 7
64
Estimating the Degradation Function
Three principal ways to estimate the degradation function
1. Observation
2. Experimentation
3. Mathematical Modeling
Lecture # 7
65
Mathematical Modeling (1)
Environmental conditions cause degradation

A model about atmospheric turbulence
k ( u 2 v 2 )5/6
H (u, v) e
k : a constant that depends on
the nature of the turbulence
Lecture # 7
66
11/5/2015
Lecture # 7
67
Image deblurring techniques
Goal: to process an image that has been subject to

blurring caused, e.g., by camera motion during
image capture or poor focusing of the lenses.
Simplest technique: inverse filtering
where: 1/H(u, v) is the FT of the restoration filter
Problems:
Division by zero
0/0 indetermination
If there is also noise, it will heavily influence the
calculations
Lecture # 7
68

Inverse
filtering
Division by zero problem has 2 possible solutions:

Apply a (Butterworth) lowpass filter with transfer function L(u,v) to
the division, thus limiting the restoration to a range of frequencies
below the restoration cutoff frequency:
Use constrained division, where a threshold value T is chosen such

that, if |H(u,v)| < T, the division does not take place and the original
value is kept untouched, i.e.:
Lecture # 7
69
Image deblurring
techniques
FIGURE: Example of image

restoration using inverse
filtering: (a) input (blurry)
image; (b) result of naive
inverse filtering; (c) applying a
10th-order Butterworth lowpass filter with cutoff frequency
of 20 to the division; (d) same
as (c), but with cutoff
frequency of 50; (e) results of
using constrained division, with
threshold T = 0.01; (f) same
as (e), but with threshold T =
0.001.
Lecture # 7
70
Wiener filtering (1942)

Image restoration solution that can be applied to images that have been
subject to a degradation function and also contain noise (worst-case
scenario).
Attempt to model the error in the restored image through statistical
methods, particularly the minimum mean-square estimator: once the error
is modeled, the average error is mathematically minimized.
The transfer function of a Wiener filter is given by:
where H(u, v) is the degradation function and K is a constant used to approximate the
amount of noise.
Lecture # 7
72
FIGURE: Example of
image restoration
using Wiener filtering:
(a) input image
(blurry and noisy); (b)
result of inverse
filtering, applying a
10th-order
Butterworth low-pass
filter with cutoff
frequency of 50 to the
division; (c) results of
Wiener filter, with K =
103; (d) same as
(c), but with K = 0.1.
Lecture # 7
73
FIGURE: Example of
image restoration
using Wiener filtering:
(a) input (blurry)
image; (b) result of
inverse filtering,
applying a 10th-order
Butterworth low-pass
filter with cutoff
frequency of 50 to the
division; (c) results of
Wiener filter, with K =
105; (d) same as
(c), but with K = 0.1.
Lecture # 7
74

Lect07 Image Restoration

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Lect07 Image Restoration

Hochgeladen von

Copyright:

Verfügbare Formate

Digital Image Processing

Noise in the context of image processing

Noise removal techniques are typically used in image

Deblurring techniques are typically used in image

Digital Image Processing

Image restoration: recover an image that has been

Model the degradation and applying the inverse process in

Digital Image Processing

A Model of Image Degradation/Restoration

Digital Image Processing

A Model of Image Degradation/Restoration

position-invariant process, then

the degraded image is given in the spatial domain by

Digital Image Processing

A Model of Image Degradation/Restoration

The model of the degraded image is given in

Digital Image Processing

The principal sources of noise in digital images arise during

Digital Image Processing

Noise Models (1)

Erlang (gamma) noise: Laser imaging

Uniform noise: Least descriptive;

Impulse noise: quick transients,

Basis for numerous random number

such as faulty switching

Digital Image Processing

Digital Image Processing

Gaussian Noise (1)

Digital Image Processing

Gaussian Noise (2)

Digital Image Processing

Erlang (Gamma) Noise

The mean and variance of this density are given by

Digital Image Processing

The mean and variance of this density are given by

Digital Image Processing

The mean and variance of this density are given by

Digital Image Processing

Impulse (Salt-and-Pepper) Noise

if b a, gray-level b will appear as a light dot,

Examples of Noise: Original Image

Digital Image Processing

Examples of Noise: Noisy Images(1)

Digital Image Processing

Examples of Noise: Noisy Images(2)

Digital Image Processing

Digital Image Processing

Periodic noise in an image arises typically from electrical or

It is a type of spatially dependent noise

Periodic noise can be reduced significantly via frequency

Digital Image Processing

An Example of Periodic Noise

Digital Image Processing

Digital Image Processing

Estimation of Noise Parameters (1)

Digital Image Processing

Estimation of Noise Parameters (2)

Digital Image Processing

Restoration in the Presence of Noise Only

Digital Image Processing

Spatial Filtering: Mean Filters (1)

Arithmetic mean filter

Digital Image Processing

Spatial Filtering: Mean Filters (2)