Digital Image Processing: Image Restoration

CSC447: Digital Image
Processing
Chapter 5: Image Restoration
Prof. Dr. Mostafa Gadal-Haqq M. Mostafa
Computer Science Department
College of Computer & Information Sciences
AIN SHAMS UNIVERSITY

Contents
In this lecture we will look at image restoration
techniques used for noise removal
 What is image restoration?
 Noise and images
 Noise models
 Noise removal using spatial domain filtering
 Periodic noise
 Noise removal using frequency domain
filtering
CS447: Introduction to Digital Image Processing Prof. Dr. Mostafa GadalHaqq. 2

What is Image Restoration?
Image restoration attempts to restore images
that have been degraded
 Identify the degradation process and attempt
to reverse it
 Similar to image enhancement, but more
objective

Noise and Images
The sources of noise in digital images arise
during image acquisition (digitization) and
transmission
 Imaging sensors can be affected by ambient
conditions
 Interference can be added
to an image during transmission

Noise Model
We can consider a noisy image to be modelled
as follows:
where f(x, y) is the original image pixel, η(x, y) is
the noise term and g(x, y) is the resulting noisy
pixel
If we can estimate the model of the noise in an
image, this will help us to figure out how to
restore the image.
),(),(),( yxyxfyxg 

Noise Models
Gaussian Rayleigh
Erlang Exponential
Uniform
Impulse
There are many different
models for the image
noise term η(x, y):
 Gaussian
 Most common model
 Rayleigh
 Erlang
 Exponential
 Uniform
 Impulse
 Salt and pepper noise

Noise Example
The test pattern to the right is ideal for
demonstrating the addition of noise
The following slides will show the result of adding
noise based on various models to this image
Image Histogram

Noise Example (cont…)
Gaussian Rayleigh Erlang

Noise Example (cont…)
Exponential Uniform Impulse

Filtering to Remove Noise
We can use spatial filters of different kinds to
remove different kinds of noise
The arithmetic mean filter is a very simple one
and is calculated as follows:
This is implemented as the
simple smoothing filter
Blurs the image to remove
noise


xySts
tsg
mn
yxf
),(
),(
1
),(ˆ
1/9
1/9
1/9
1/9
1/9
1/9
1/9
1/9
1/9

Other Means
There are different kinds of mean filters all of
which exhibit slightly different behaviour:
 Geometric Mean
 Harmonic Mean
 Contra-harmonic Mean

Other Means (cont…)
There are other variants on the mean which can
give different performance
Geometric Mean:
 Achieves similar smoothing to the arithmetic
mean, but tends to lose less image detail
mn
Sts xy
tsgyxf
1
),(
),(),(ˆ








 

Harmonic Mean:
 Works well for salt noise, but fails for pepper
noise.
 Also does well for other kinds of noise such
as Gaussian noise


xySts tsg
mn
yxf
),( ),(
1
),(ˆ

Contra-harmonic Mean:
 Q is the order of the filter and adjusting its
value changes the filter’s behaviour
 Positive values of Q eliminate pepper noise
 Negative values of Q eliminate salt noise






xy
xy
Sts
Q
Sts
Q
tsg
tsg
yxf
),(
),(
1
),(
),(
),(ˆ

Noise Removal Examples
Original
Image
Image
Corrupted
By Gaussian
Noise
After A 3*3
Geometric
Mean Filter
After A 3*3
Arithmetic
Mean Filter

(cont…)
Image
Corrupted
By Pepper
Noise
Result of
Filtering Above
With 3*3
Contraharmonic
Q=1.5

(cont…)
Image
Corrupted
By Salt
Noise
Result of
Filtering Above
With 3*3
Contraharmonic
Q=-1.5

Contra-harmonic Filter
Choosing the wrong value for Q when using the
contra-harmonic filter can have drastic results

Order Statistics Filters
Spatial filters that are based on ordering the
pixel values that make up the neighbourhood
operated on by the filter
Useful spatial filters include
 Median filter
 Max and Min filter
 Midpoint filter
 Alpha trimmed mean filter

Median Filter
Median Filter:
 Excellent at noise removal, without the
smoothing effects that can occur with other
smoothing filters
 Particularly good when salt and pepper noise
is present
)},({),(ˆ
),(
tsgmedianyxf
xySts 


Max and Min Filter
Max Filter:
Min Filter:
Max filter is good for pepper noise and Min is
good for salt noise
)},({max),(ˆ
),(
tsgyxf
xySts 

)},({min),(ˆ
),(
tsgyxf
xySts 


Midpoint Filter
Midpoint Filter:
Good for random Gaussian and uniform noise



 

)},({min)},({max
2
1
),(ˆ
),(),(
tsgtsgyxf
xyxy StsSts

Alpha-Trimmed Mean Filter
Alpha-Trimmed Mean Filter:
 We can delete the d/2 lowest and d/2 highest
grey levels
 So gr(s, t) represents the remaining mn – d
pixels


xySts
r tsg
dmn
yxf
),(
),(
1
),(ˆ

Image
Corrupted
By Salt And
Pepper Noise
Result of 1
Pass With A
3*3 Median
Filter
Result of 2
Passes With
A 3*3 Median
Filter
Result of 3
Passes With
A 3*3 Median
Filter

(cont…)
Image
Corrupted
By Pepper
Noise
Image
Corrupted
By Salt
Noise
Result Of
Filtering
Above
With A 3*3
Min Filter
Result Of
Filtering
Above
With A 3*3
Max Filter

(cont…)
Image
Corrupted
By Uniform
Noise
Image Further
Corrupted
By Salt and
Pepper Noise
Filtered By
5*5 Arithmetic
Mean Filter
Filtered By
5*5 Median
Filter
Filtered By
5*5 Geometric
Mean Filter
Filtered By
5*5 Alpha-Trimmed
Mean Filter

Periodic Noise
 Typically arises due to
electrical or electromagnetic
interference
 Gives rise to regular noise
patterns in an image
 Frequency domain
techniques in the Fourier
domain are most effective at
removing periodic noise

Band Reject Filters
 Removing periodic noise form an image involves
removing a particular range of frequencies from
that image
 An ideal band reject filter is given as follows:












2
),(1
2
),(
2
0
2
),(1
),(
0
00
0
W
DvuDif
W
DvuD
W
Dif
W
DvuDif
vuH

Band Reject Filters (cont…)
The ideal band reject filter is shown below,
along with Butterworth and Gaussian versions
of the filter
Ideal Band
Reject Filter
Butterworth
Band Reject
Filter (of order 1)
Gaussian
Band Reject
Filter

Band Reject Filter Example
Image corrupted by
sinusoidal noise
Fourier spectrum of
corrupted image
Butterworth band
reject filter
Filtered image

Adaptive Filters
 The filters discussed so far are applied to an
entire image without any regard for how
image characteristics vary from one point to
another
 The behaviour of adaptive filters changes
depending on the characteristics of the image
inside the filter region
 We will take a look at the adaptive median
filter

Adaptive Median Filtering
 The median filter performs relatively well on
impulse noise as long as the spatial density
of the impulse noise is not large
 The adaptive median filter can handle much
more spatially dense impulse noise, and also
performs some smoothing for non-impulse
noise
 The key element in the adaptive median filter
is that the filter size changes depending on
the characteristics of the image

Adaptive Median Filtering (cont…)
Remember that filtering looks at each original pixel
image in turn and generates a new filtered pixel
First examine the following notation:
 zmin = minimum grey level in Sxy
 zmax = maximum grey level in Sxy
 zmed = median of grey levels in Sxy
 zxy = grey level at coordinates (x, y)
 Smax =maximum allowed size of Sxy

Level A: A1 = zmed – zmin
A2 = zmed – zmax
If A1 > 0 and A2 < 0, Go to level B
Else increase the window size
If window size ≤ Smax repeat level A
Else output zmed
Level B: B1 = zxy – zmin
B2 = zxy – zmax
If B1 > 0 and B2 < 0, output zxy
Else output zmed

The key to understanding the algorithm is to
remember that the adaptive median filter has
three purposes:
 Remove impulse noise
 Provide smoothing of other noise
 Reduce distortion

Adaptive Filtering Example
Image corrupted by salt
and pepper noise with
probabilities
Pa = Pb=0.25
Result of filtering with
a 7 * 7 median filter
Result of adaptive
median filtering with
i = 7

Summary
 In this lecture we studied image restoration for
noise removal
 Restoration is slightly more objective than
enhancement
 Spatial domain techniques are particularly useful
for removing random noise
 Frequency domain techniques are particularly
useful for removing periodic noise
 Behaviour of adaptive filters changes depending
on the characteristics of the image inside the filter
region

HW 4
 Solve problems 5.1 to 5.12.

Digital Image Processing: Image Restoration

More Related Content

What's hot (20)

Similar to Digital Image Processing: Image Restoration (20)

More from Mostafa G. M. Mostafa (20)

Recently uploaded (20)

Digital Image Processing: Image Restoration