Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 8

BCS-V

Submitted by: Shahbano (039),


Group-B
Sumbal Naz (040)
Submitted to: Ms. Mehmoona
khanum

Image
Enhancement
Theory
of Automata
Research Paper

Study of Image Enhancement Techniques for Grayscale Images


Sumbal Naz1, Shahbano2, Memoona Khanam3
123

Computer Science Department, Fatima Jinnah Women University

Abstract
The presented paper discuss different techniques of
image enhancement like histogram equalization
which is explained in detail with examples and
sample data of 8x8 matrix is also processed, point
processing techniques which includes linear,
logarithmic and power law transformations. Along
with the description of these techniques code is also
implemented in MATLAB with the help of general
forms and the results of the processed images are
shown in the form of enhanced images. These
techniques are used to enhance different types of
images like some are used to enhance satellite
images, some are used for x-ray images and some are
used for dark images. Our main emphasisis on gray
scale images.
Keywords:
Image
enhancement,
Histogram
equalization (HE), spatial domain, Gray Level
Transformation, Linear Transformation, Logarithmic
transformation, Power law.
1.

Introduction

Image processing has a wide range of applications in


computer vision, multimedia communication,
television broadcasting, etc. that demand very good
quality of images. The quality of an image degrades
due to introduction of noise during acquisition,
transmission/ reception and storage/ retrieval
processes. An image with high contrast and
brightness is called fine quality image while a poor
quality image is defined by low contrast and poorly
defined boundaries between the edges. For better
understandability of image we use Image
enhancement technique refers to sharpening of image
features such as edges, boundaries or contrast to
make it more useful. The fundamental goal of image
enhancement is to process the input image in such a
way that the output image is more suitable for
interpretation by the humans as well as by machines.
Usually two types of images are enhanced, hazy
images and t color images. Image enhancement is the
most interesting and the visually appealing areas of
image processing. Techniques used for grayscale
image enhancement are histogram stretch, histogram
equalization and adaptive contrast enhancement.

However, our main emphasis is on the histogram


equalization. [1]
One technique cannot be used for all type of images,
for this reason image enhancement approaches a
major category which is spatial domain. It refers to
the direct manipulation of pixel in an image. In
spatial domain there are three basic gray level
transformations, which are linear, logarithmic and
Power-law [1].
Techniques are implemented using Matrix
Laboratory (MATLAB) which provide ability to
write the algorithm in a high-level programming
language with built in visualization tools.
2.

Literature review

In this section we study techniques which are already


reported like histogram equalization and spatial
domain. Histogram equalization is an image
enhancement technique. It is simple and popular
technique but still it has some limitations. But due to
these limitations several histogram equalization
methods has been developed. Histogram equalization
causes the mean luminance of the image, produces
artifacts and unnatural enhancements and does not
consider local information in its process. We also
define the other processing techniques of Image
enhancement. Thus the contribution of this paper is to
classify and review image enhancement processing
techniques. [2]
Contrast enhancement is considered as an
optimization problem that minimizes the cost.
Histogram equalization is an efficient technique for
contrast enhancement. However, conventional
histogram equalization (HE) typically results in
excessive contrast enhancement, which in turn gives
the processed image. But by introducing special
terms of the histogram equalization the level of
contrast enhancement can be adjusted, noise
robustness; stretching and brightness can also be
improved to a high level. [3]
In writing, many efficient digital image filters are
found that perform well under low noise conditions.
But their performance is not so good under fair and
high noise conditions. Thus, it is felt to develop well-

organized but simple algorithms to restrain modest


and high power noise in an image. Noise in digital
images is found to be additive in nature. Such a noise
is referred to as Additive White Gaussian Noise
(AWGN). Image de-noising is usually essential to be
performed in order to do segmentation, feature
extraction, object recognition, texture analysis, etc.
The purpose of de-noising is to suppress the noise
quite effectively while smoothing the edges and the
other detailed features as much as possible. [4]
Histogram Equalization (HE) is a simple and
effective image enhancement technique. But, it tends
to change the mean brightness of the image, but it
does not work well for consumer product. To
preserve brightness and to enhance contrast of
images, there are many methods which are to be
introduced, but many of them present unwanted
artifacts such as intensity saturation, to increase the
enhancement and to amplify the noise. Available
histogram equalization methods are reviewed and
compared with other methods of image processing in
order to contrast the brightness and other areas of the
image, and also to evaluate contrast enhancement. [5]
Image Enhancement is one of the most essential and
complicated techniques in image research. The aim of
image enhancement is to recover the visual form of
an image, or to provide a pleasing effect in an image.
Many images like medical images, satellite images,
aerial images and even real life photographs suffer
from poor contrast and noise. In this way it is
required to enhance the image and to improve the
quality so that it gives the pleasing effect. Image
Enhancement techniques in medical images detection
and analysis, remove blurring and noise, increasing
contrast, and revealing details. The existing
techniques of image enhancement can be classified
into two categories: Spatial Domain and Frequency
domain enhancement. In this paper, we present an
overview of image enhancement processing
techniques in spatial domain. [6]
The main purpose of image enhancement is to
practice an image so that result is more pleasing than
the original image. Digital image processing
techniques has two domains: spatial domain and
frequency domain. Image enhancement plays a vital
role in every field where images have to be
understood and analyzed. Many images like medical
images, satellite images, microscopy images, aerial
images and even real life photographs go through
from poor contrast and noise. It is necessary to
enhance the contrast and remove the noise to increase
image visual quality. This paper focuses on different

image enhancement techniques and better approach


for future research. [7]

3.

Histogram Equalization

Histogram Equalization is a technique where the


histogram of the resultant image is as flat as possible.
The theoretical basis for histogram equalization
involves probability theory, where the histogram is
treated as the probability distribution of the gray
levels. [1]
Assuming X to be an image whose pixel value is
accessed as X (i,j). It is composed of L discrete gray
levels denoted by {X0,X1,,XL-1}. Here, X(i,j)
represents the intensity of the image at spatial
location (i,j) with the condition that X(i,j)E{ X0,X1,
,XL-1}. As the intensities are all discrete values,
the histogram of a digital image is a discrete function
[2].
Histogram h is defined as

h(Xk)=nk,
for k=0, 1, , L-1

Where Xk is the k-th gray level and nk presents the


number of times that the gray level X k appears in the
image. In other words, the histogram is the frequency
of occurrence of the gray levels in the image [2].
Hand calculation for better understandability are
performed. Table 1 contain randomly generated data
for 8x8 array to represent and image. Figure 1 shows
Original image of pixel values in Table 1.
Table 1: Original data of 8x8 array of 3-bit image
0
0
3
2
7
6
6
6

0
2
3
3
7
4
6
7

2
2
0
1
6
7
5
5

3
1
3
2
2
1
2
5

5
3
4
3
6
7
2
2

3
1
5
3
4
1
1
1

2
0
0
4
6
5
1
0

3
2
1
6
5
4
0
0

Figure 1: 3-bit image of 8x8 array

Figure 2: 3-bit Histogram Equalized image

Table 2 shows calculations. Step # 1 include number


of each pixel. In Step # 2 Running Sum is calculated.
Step # 3 shows normalization and then multiplication
by highest intensity value i.e. 7. If the image is 8 bit,
multiply by 255. Table 3 consists of Step # 4 which
shows new image. Step # 6 number of each gray level
of new image. Figure 2 shows new 3-bit image
obtained after calculations.
Table 2: Results of Calculations for 8x8 array

Gray level
No. of
pixels

0
9

Running
Sum

Normalizi
ng and
multiplyin
g by 7

1
1
4
3
7
6
6
6

1
3
4
4
7
5
6
7

Gray
Level
No. of
pixels

0
0

Step # 1
1 2 3
9 1 1
1 0
Step # 2
1 2 3
8 9 9
Step # 3
2 3 4

4
5

5
7

6
8

7
5

4
4

5
1

5
9

6
4

Table 3: New 3-bit image


Step # 4
3
4
6
4
3
2
4
2
1
4
5
6
2
3
4
4
6
3
6
5
7
2
7
2
6
3
3
2
6
6
3
2
Step # 5
1 2 3
4
5
9

11

10

Figure 3: Histogram of Original 3-bit image

Figure 4: Histogram of New 3-bit image


3
1
1
5
6
6
2
1

4
3
2
6
6
5
1
1
6

15

Figure 3 shows the histogram of original image


Figure 4 shows Equalized image of the data given in
Table 1, Table 2 and Table 3. Advantage of histogram
equalization is that it is easy to use and
straightforward technique. On the other hand it may
increase the contrast of background noise. [3] Figure
5 is an example of noisy image produced after
histogram equalization.

Figure 5: (a) Orignal Image (b) Equalized Image


Table 4: Advantages and disadvantages
Advantages
It is a fairly straightforward technique and an invertible
operator. So in theory, if the histogram equalization
function is known, then the original histogram can be
recovered. [3]
Histogram equalization is a simple and effective
contrast enhancement technique which distributes pixel
values uniformly such that enhanced image have linear
cumulative histogram. [5]
It stretches the contrast of the high histogram regions
and compresses the contrast of the low histogram
regions.[5]

Figure 8: Equalized Satellite Image


Figure 9 and Figure 10 shows Histograms of the
images displayed in figure 7 and figure 8.

Code for histogram equalization in MATLAB


evaluates results shown in Figure 6:
Figure 9: Histogram Equalization (a) Original
Image (b) Equalized Images

Figure 10: Bar graph of (a) Original Image (b)


Equalized Image
Figure 6: (a) Grayscale Image (b) Equalized
Image
Histogram equalization is better used for
enhancement of satellite images. So a Satellite image
which is shown in Figure 7 is equalized and new
image is shown in Figure 8.

Figure 7: Satellite image

Figure 11 displays flowchart of the steps involved in


histogram equalization.

Negative/Identity: This transformation reverses the


gray level order. For L gray levels, the transformation
has the form:
s = (L-1) r
Negative images are useful for enhancing white or
grey detail embedded in dark regions of an image. [1]
[7]Code for Negative transformation in Matlab
evaluated results which are shown in Figure 12.

Figure 12: (a) X-ray image (b) Negative of X-ray


image
Figure 11: Flowchart for Histogram Equalization
4.

Spatial Domain

Most spatial domain enhancement operations can be


generalized as:
g(x,y)=T[f(x,y)].
Where, f (x, y) = the input image, g (x, y) = the
processed/output image and T = some operator
defined over some neighborhood of (x, y). Spatial
domain techniques directly deal with image pixels.
Spatial techniques are particularly useful for directly
altering the gray level values of pixels and hence the
overall contrast of the image. But the image is
enhance in a uniform manner which sometimes
produced undesirable results. Point processing
operations are the simplest spatial domain operations.
Few of them are Linear, Logarithmic and Power law
transformations. Pixel values of the processed image
depend upon the pixel values of original image. The
expression can be stated as g(x,y) = T[f(x,y)], where
T is gray level transformation in point processing. [6]

Thresholding Transformation: it is achieved in a


normalized gray scale as pixel values of threshold
image are either 0s or 1s. If r > threshold then s =
1.0 and if r <= threshold then s= 0.0 where, r is the
pixel of original image and s is the pixel of
transformed image. Resultant image is also called
binary image. These are useful in image segmentation
to isolate an image of interest from back ground. [7].
Code of Thresholding Transformation in MATLAB
evaluates resultsshown in Figure 13with thresholding
value as 100.

Figure 13: (a) Grayscale Image of Earth (b)


Threshold image of earth

4.2. Log Transformation:


4.1. Linear Transformation

Log Transformation Technique can be used for


contrast enhancements of dark images. The general
form is represented as

s = c log (1 + r)
Where, s is the output pixel value, r is the input pixel
value and c is a constant. [1][7][8]. Code for Log
Transformation in MATLAB evaluates results shown
in Figure 14:

5.

Analysis of Techniques
Table 5: Parameters

Parameter
Efficiency

Performance

Availability
Figure 14: (a) Original Image (b) Transformed
Image

Reliability

4.3. Power law transformation:

Safety

Power law transformation technique is commonly


used gray level transformation, in which narrow
range of input value are mapped into wider range of
output values. Hence increasing the contrast. It can
be represented as

Security

Noise
Sharpness

s = br
Where, s is the output pixel value, r is the input pixel
value, b is a scaling constant and is the power to
which the input gray level is raised. Advantage is of
this transformation is that it is possible to control the
transformation function by varying the parameter .
[1] [7] [8] Code of Power law transformation
evaluates results shown in Figure 15.
Parameters
Efficiency
Performance
Availability
Reliability

Sharpness

Contrast

Description
It signifies a level of performance
that describes a process that uses the
lowest amount of inputs to create
the greatest amount of outputs.
The accomplishment of a given task
measured against preset known
standards of accuracy, completeness,
cost, and speed.
Ability of technique to deliver
services when required
Ability of technique to deliver
services as specified
Ability of technique to operate
without failure
Ability of technique to protect itself
against accidental or deliberate
intrusion
Undesirable by-product of image
that adds extraneous information
Subjective quality of an image
indicating clear or distinct
reproduction of detail associated
with resolution and contrast
Separation between the darkest and
brightest areas of the image

Very go

Very go

Very go
good
Good, b

Very go

Good ,

Very go
good

Good, n

Table 6: Parametric evaluation of Image


Enhancement Techniques
Histogram
Equalization
Very good
Very good
Good
Good
Very good
Good
Not good
Good
Not good

Linear
Transformation
Very good
Very good
Good
Very Good
Very good
Good
Good
Good
Good

Logarithmic
Transformation
Good
Good
Good
Not good
Very good
Good
Good
Not good
Not good

Power Law
Transformation
good
Very good
Good
Good
Very good
Good
Good
Good
Good

Conclusion
Figure 15: (a) Original Image, (b) Image with
=0.2, (c) Image with =0.4, (d) Image with =0.6

Possible
Very go

Digital image processing is the use of computer


algorithms to perform image processing on digital

images. Basically image enhancement is the process


of adjusting the digital images so that the resultant
images produced are more suitable to display and
have a more visually pleasing effect. Histogram
equalization is best suitable for land images taken
from satellite, radar images and texture synthesis. Its
code can be used to equalize different images but can
also produce noisy images and the histogram of
images show balance in the number of pixels used.
Linear transformation can be used for producing
negative of images and black and white images. Log
transformation is used for finding details and Power
law can produce same image with different contrasts.
This research paper provides an overview of few of
image enhancement techniques and their comparative
analysis.

3.

4.

5.

Future Work
As the requirements change with respect to time so in
future we need more efficient algorithms. The
techniques discussed above are good for image
enhancement but they have some flaws. Histogram
equalization has equalize the image but it does not
reduce the noise of an image due to this the image
seems blur so in future different algorithms will be
implemented in order to reduce noise. The above
technique enhances only the low contrast images, in
future, implementations will be done for high
contrast images. Image enhancement techniques
mention in this paper are all for the gray scale
images, such techniques can also be implemented for
color images. All the above facts involves a lot of
calculations and they are not practical for 3-D
medical image enhancement so work should be done
for 3-D images.

6.

7.

References
1.
2.

Rafael C. Gonzalez, Richard E. Woods. Digital


Image Processing Third Edition
Nicholas Sia Pik Kong, Haidi Ibrahim, and Seng
Chun Hoo A Literature Review on Histogram
Equalization and ItsVariations for Digital Image
Enhancement.International Journal of
Innovation, Management and Technology, Vol. 4,
No. 4, August 2013. Available:
http://www.ijimt.org/index.php?
m=content&c=index&a=show&catid=46&id=70
7

8.

Manvi, Rajdeep Singh Chauhan, Manpreet


Singh, Image Contrast Enhancement Using
Histogram Equalization International Journal
of Computing & Business Research. Available:
http://www.researchmanuscripts.com/isociety201
2/33.pdf
NILAMANI BHOI. Development of Some
Novel Spatial-Domain and Transform Domain
Digital Image Filters in proceeding National
Institute of Technology, Rourkela, INDIA
January 2009. Available at:
http://ethesis.nitrkl.ac.in/1529/1/PhD_Thesisbhoi.pdf
Raju. A , Dwarakish. G. S and D. Venkat Reddy.
A Comparative Analysis of Histogram
Equalization based Techniques for Contrast
Enhancement and Brightness Preserving,
International Journal of Signal Processing, Image
Processing and Pattern Recognition Vol.6, No.5
(2013). Available at:
http://www.sersc.org/journals/IJSIP/vol6_no5/31
.pdf
Dr. S. S. Bedi, Rati Khandelwal. Various Image
Enhancement Techniques- A Critical Review.
International Journal of Advanced Research in
Computer and Communication EngineeringVol.
2, Issue 3, March 2013. Available:
http://www.ijarcce.com/upload/2013/march/55rati%20-%20various%20image
%20enhancement.pdf
Suneetha, Dr. T. Venkateswarlu, Enhancement
Techniques for Gray scale Images in Spatial
Domain International Journal of Emerging
Technology and Advanced Engineering Website:
www.ijetae.com (ISSN 2250-2459, Volume 2,
Issue 4, April 2012). Available:
http://www.ijetae.com/files/Volume2Issue4/IJET
AE_0412_03.pdf
Snehal O.Mundhada, V. K.Shandilya. Spatial
and Transformation Domain Techniques for
Image Enhancement International Journal of
Engineering Science and Innovative Technology
(IJESIT) Volume 1, Issue 2, November 2012.
Available: http://www.ijesit.com/Volume
%201/Issue%202/IJESIT201211_32.pdf

You might also like