UNIT-3

IMAGE RESTORATION

Image Restoration attempts to reconstruct an image that has been degraded by using a
priori knowledge of the degradation phenomenon.

Restoration involves modeling of degradation and then applying the inverse process in
order to recover the original image.

Image enhancement Vs image Restoration

Enhancement Restoration

Enhancement is s process that aims to Restoration is the process of inverting a

improve image ‘bad’ quality image so degradation using knowledge about its
that it looks better nature

It is subjective process It is a objective process

It gives better visual representation It removes effects of sensing environment

Enhancement techniques: Restoration techniques:
Contrast stretching etc. Inverse filtering , wiener filtering

No model required Mathematical model of degradation is

needed

Model for image Degradation/Restoration Process.

To restore a degraded image to its original content and quality

1
  The degraded image in the spatial domain is given by

Where, h(x,y) is spatial representation of degradation function and symbol *

represents convolution.

 The degraded image in frequency domain is given by

G(u,v)=F(u,v)H(u,v)+N(u,v)

Noise Models

The principal source of noise in digital images arises during image acquisition and /or
transmission.
The performance of imaging sensors is affected by a variety of factors, such as
environmental conditions during image. by the quality of the sensing elements
themselves. Images are corrupted during transmission principally due to interference
in the channels used for transmission.

Types of noise:

1. Gaussian Noise (normal noise)

These noise models are used frequently in practices because of its tractability in both
Spatial and frequency domain

The PDF of Gaussian random variable, z is given by

2
2. Rayleigh Noise:
Unlike Gaussian distribution, the Rayleigh distribution is no symmetric. It is given by
the formula

3. Gamma Noise

3
4. Impulse (salt-and-pepper) Random Noise

a is pepper, dark
b is salt, which is brighter

4
Restoration in the presence of Noise only- Spatial filtering:

When the only degradation present in an image is noise, i.e

g(x,y)=f(x,y)+η(x,y)
or
G(u,v)= F(u,v)+ N(u,v)

Spatial filtering can be done when only additive noise is present. The following
techniques can be used to reduce the noise effect:

Mean Filter

(a) Arithmetic Mean Filter

 This filter removes local variations in image Noise is reduced as a result
of blurring
 It is useful in removing Gaussian Noise and uniform noise

Using the pixels in the region defined by Sxy

(b) Geometric Mean filters:

 This filter eliminate Gaussian Noise

 It is ineffective for paper type noise
 An image restored using a geometric mean filter is given by the
expression

5
(c)Harmonic mean filters:-

2. Order statistics filter:

Order statistics filters are spatial filters whose response is based on ordering the pixel
contained in the image area encompassed by the filter. The response of the filter at any
point is determined by the ranking result.

i. Median filter:-

 It replaces the value of a pixel by the median of gray levels in the

Neighborhood of the pixel
 Median filter provide excellent noise reduction with less blurring
 Median filters are effective for bipolar and unipolar impulse noise.

ii. Min and Max filter:-

Max filter:-


 It select the Maximum value in the sorted list

Min filter:-
 It is used for finding the darkest point in an image
 It select the minimum value in the sorted list

6
iii. Midpoint Filter:-

The midpoint filter simply computes the midpoint between the maximum and
minimum values in the area encompassed by the filter

Iv Alpha-trimmed mean filter:-

It is useful in situation including multiple types of noise, such as combination of salt-

and-pepper and Gaussian Noise

 If d=0, arithmetic mean filter

 If d= mn-1, median filter

Periodic Noise By Frequency Domain Filtering

These types of filters are used for this purpose-
1. Band Reject Filters
 It removes a band of frequencies above the origin of the Fourier
transformer
 Band Reject filters are best suitable for reducing Periodic noise

7
 1.1 Ideal Band reject Filter

D(u,v)- the distance from the origin of the centered frequency band.
W- the width of the band
Do- the radial center of the frequency band

1.2Butterworth Band reject Filter

1.3 Gaussian Band rejects Filter

2. Band Pass Filters:-

 Band pass filters are the exact opposite of band reject filters. They pass
a band of frequencies, around some frequency, say D0 (rejecting the rest)
 One can write:

 Band pass filter is usually used to isolate components of an image that

correspond to a band of frequencies.

8
Notch Filters
 A Notch filters reject (or passes) frequencies in predefined neighbourhoods
above a center frequency.

9
Inverse Filters:

The simplest approach to restoration is direct inverse filtering. This is obtained as

It shows an interesting result that even if we know the depredation function we cannot
recover the underrated image exactly because N(u,v) is not known .
If the degradation value has zero or very small values then the ratio N(u,v)/H(u,v)
could easily dominate the estimate F(u,v).

The inverse filter works fine provided there is no noise

Minimum Mean Square Error (Wiener) Filtering

 It is also known as Minimum Mean Square Error Filter or Least Square

Error Filter
 The wiener filtering approach uses degradation function and statistical behavior
of noise into the restoration process.


10
It is assumed that
1. the noise and the image are uncorrelated one or other has zero mean values
2. The gray levels in the estimate are a linear function of the levels in the
degraded image.

Based on this assumption the minimum error function of the above expression is given

11
 Important Questions of UNIT-3

1. What is Image Restoration?

2. Differentiate between image enhancement and restoration
3. What is Image Restoration? Draw and explain the basic block diagram of the
restoration process.
4. Explain Image degradation/ Restoration Process. Explain all noises with their
PDF.
5. List the properties of degradation model.
6. What do you mean by Gaussian Noise and why averaging filter is used to
eliminate it?
7. Discuss order statics filters with suitable example.
8. Discuss the process of image smoothing using ideal low pass filters and
Butterworth low pass filters
9. Compare the inverse filter with Wiener filter.
10. Explain Band pass Filter Technique for noise reduction..
11. Write short note on:
I. Band pass filter technique for noise reduction
II. Minimum error square filtering.
12. Explain why Band Reject filters are best suitable for reducing Periodic noise.
Explain all Band Reject filters in detail. Obtain corresponding epression for
Band Pass filters.

13. Wiener filter with SNR, MSE ratio for spatial and frequency domain.

14. What is image restoration? Explain in detail the image restoration in presence
of noise only.

15. Explain the process of minimum mean square error restoration.

16. Explain the concept of Inverse Filtering also mention the limitations of it.

12
13