Chapter five
Image Restoration and Reconstruction
Content
Introduction
Models of images and noise
Estimation of noise parameters
Restoration in the presence of noise
Introduction
The principal goal of restoration techniques is to improve an image in some predefined
sense.
Image restoration is to recover an image that has been degraded by using a priori
knowledge of the degradation phenomenon.
Image enhancement vs image restoration
❖ Enhancement is for vision
❖ Restoration is to recover the original image
Image restored is an approximation of the original image.
❖ Criteria for the goodness
Cont…
A Model of the Image Degradation/Restoration Process
Cont….
Noise Models
The principal sources of noise in digital images arise during image acquisition and/or
transmission.
Spatial and Frequency Properties of Noise In the spatial domain, we are interested in
the parameters that define the spatial characteristics of noise, and whether the noise is
correlated with the image.
Frequency properties refer to the frequency content of noise in the Fourier sense.
Cont…
Some Important Noise Probability Density Functions
Gaussian noise
Because of its mathematical tractability in both the spatial and frequency domains,
Gaussian (normal) noise models are used frequently in practice.
The probability density function (PDF) of a Gaussian random variable, z , is given by
Cont…
Rayleigh noise
The probability density function of Rayleigh noise is given by
Cont…
Erlang (gamma) noise
The probability density function of Erlang noise is given by
Cont…
Exponential noise
The PDF of exponential noise is given by
Cont…
Uniform noise
The PDF of uniform noise is given by:
Cont…
Impulse (salt-and-pepper) noise
The PDF of impulse noise is given by
Cont…
Cont…
Add spatial noise to an image (Noisy images and their histograms)
Cont…
Cont…
Cont…
Periodic Noise
Periodic noise in an image arises typically from
electrical or electromechanical interference
during image acquisition.
The periodic noise can be reduced significantly
via frequency domain filtering
Estimation of Noise Parameters
The parameters of periodic noise can be estimated by inspection of the Fourier spectrum of the
image.
Periodic noise tends to produce frequency spikes, which are detectable even by visual analysis.
In simplistic cases, it is also possible to infer the periodicity of noise components directly from the
image.
Automated analysis is possible if the noise spikes are either exceptionally pronounced, or when
knowledge is available about the general location of the frequency components of the interference.
It is often necessary to estimate the noise probability density functions for a particular imaging
arrangement.
Cont…
Cont…
Restoration in the Presence of Noise Only – Spatial Filtering
Cont..
Cont…
Cont…
Figure 5.7 (a) shows an 8-bit image, and Figure
5.7 (b) shows its corrupted version with additive
Gaussian noise of zero mean and variance of 400.
Figure 5.7 (c) and Figure 5.7 (d) show the result
of filtering the noisy image with a 3 x 3
arithmetic mean filter and a 3 x 3 geometric mean
filter, respectively.
Cont…
Figure 5.8 (a) and Figure 5.8 (b) show the
images corrupted by 10% pepper noise and
10% salt noise, respectively.
Figure 5.8 (c) shows the result of filtering
Figure 5.8 (a) using a contraharmonic mean
filter with Q = 1.5 .
Figure 5.8 (d) shows the result of filtering
Figure 5.8 (b) using a contraharmonic mean
filter with Q = -1.5 .
Cont…
The positive-order filter did a better job of cleaning the background, at the expense of
slightly thinning and blurring the dark areas.
The opposite was true of the negative-order filter.
In general, the arithmetic and geometric mean filters are suited for random noise like
Gaussian or uniform noise.
The contraharmonic mean filter is well suited for impulse noise, with the disadvantage that it
must known whether the noise is dark or light in order to select Q.
Cont…
Order-Statistic Filters
These are spatial filters whose response is based on ordering (ranking) the values of the pixels
contained in the image area encompassed by the filter.
Cont…
Cont…
Figure 5.10 (a) shows the image
corrupted by salt-and pepper noise with
probabilities Pa Pb = = 0.1
Figure 5.10 (b) shows the result of
median filtering with a filter of size 3 x 3
Figure 5.10 (c) and Figure 5.10 (d)
show the result of applying the same
filter on Figure 5.10 (b) and Figure 5.10
(c), respectively.
Cont…
Figure 5.11 (a) shows the result of
applying the max filter to the pepper
noise image of Figure 5.8 (a).
Figure 5.11 (b) shows the result of
applying the min filter to the image of
Figure 5.8 (b).
The min filter did a better job on noise
removal, but it removes some white
points around the border of light objects.
End of chapter five
Quizz
Discuss the difference between arithmetic mean filter, geometric
mean filter, harmonic mean filter, contra harmonic mean filter,
median filter, and also in which type of filter are used each noise
like gaussian noise, Rayleigh noise, erlang noise, exponential
noise, uniform noise, impulse (salt and pepper) noise or periodic
noise?