A Comprehensive Study of Digital photo Image Enhancement

Techniques:-

Abstract: Principle objective of digital photo Image enhancement is to process an

image so that result is more suitable than original image for specific application.
Digital image enhancement techniques provide a multitude of choices for
improving the visual quality of images. Appropriate choice of such techniques is
greatly influenced by the imaging modality, task at hand and viewing conditions.
This paper will provide an overview of underlying concepts, along with algorithms
commonly used for image enhancement. The paper focuses on spatial domain
techniques for image enhancement, with particular reference to point processing
methods and histogram processing.

I. Introduction First transferred in to frequency domain. It

Image enhancement is basically improving means that, the Fourier Transform of the
the interpretability or perception of image is computed first. All the
information in images for human viewers enhancement operations are performed on
and providing `better' input for other the Fourier transform of the image and then
automated image processing techniques. The the Inverse Fourier transform is performed
principal objective of image enhancement is to get the resultant image. These
to modify attributes of an image to make it enhancement operations are performed in
more suitable for a given task and a specific order to modify the image brightness,
observer. During this process, one or more contrast or the distribution of the grey levels.
attributes of the image are modified. The As a consequence the pixel value
choice of attributes and the way they are (intensities) of the output image will be
modified are specific toe given task. modified according to the transformation
Moreover, observer-specific factors, such as function applied on the input values.
the human visual system and the observer's
experience, will introduce a great deal of Image enhancement is applied in every field
subjectivity into the choice of image where images are ought to be understood
enhancement methods. There exist many and analyzed. For example,
techniques that can enhance a digital image Medical image analysis, analysis of images
without spoiling it. The enhancement from satellites etc. Image enhancement
methods can broadly be divided in to the simply means, transforming an
following two categories: Image f into image g using T. (Where T is
1. Spatial Domain Methods the transformation. The values of pixels in
2. Frequency Domain Methods images f and g are denoted by r and
In spatial domain techniques, we directly s,respectively. As said, the pixel values r and
deal with the image pixels. The pixel values s are related by the expression,
are manipulated to achieve desired
enhancement. In frequency domain methods, s=T(r) …….. (1)
the image is
Where T is a transformation that maps a In this paper basic image enhancement
pixel value r into a pixel value s. The results techniques have been discussed with their
of this transformation are mapped into the mathematical understanding. This paper will
grey scale range as we are dealing here only provide an overview of underlying concepts,
with grey scale digital images. So, the along with algorithms commonly used for
results are mapped back into the range [0,L- image enhancement. The paper focuses on
1], where L=2k, k being the number of bits spatial domain techniques for image
in the image being considered. So, for enhancement, with particular reference to
instance, for an 8-bit image the range of point processing methods, histogram
pixel values will be [0, 255]. processing.
I will consider only gray level images. The
same theory can be extended for the color 2. Point Processing Operation
images too. A digital gray image can have
pixel values in the range of 0 to 255. The simplest spatial domain operations
occur when the neighborhood is simply the
pixel itself. In this case T is referred to as a
grey level transformation function or a point
processing operation. Point processing
operations take the form shown in equation
(1)

i) Before image processing

Figure 2. Figure
shows basic grey level Transformations

ii) After image processing 2.1 Create Negative of an Image

Many different, often elementary and The most basic and simple operation in
heuristic methods are used to improve digital image processing is to compute the
images in some sense. The problem is, of negative of an image. The pixel gray values
course, not well defined, as there is no are inverted to compute the negative of an
objective measure for image quality. Here, image. For example, if an image of size R x
we discuss a few recipes that have shown to C, where R represents number of rows and
be useful both for the human observer C represents number of columns, is
and/or for machine recognition. These represented by I(r, c). The negative N(r, c)
methods are very problem-oriented: a of image I(r, c) can be computed as
method that works fine in one case may be N(r, c) = 255 – I(r, c) where 0 <= r <= R and
completely inadequate for another problem. 0 <= c <= C …………………………(2)
It can be seen that every pixel value from
the original image is subtracted from the
255. The resultant image becomes negative
of the original image. Negative images [3]
are useful for enhancing white or grey detail
embedded in dark regions of an image.

Figure 3 Note how much clearer the tissue is

in the negative image of the mammogram

2.2 Thresholding Transformations

Thresholding transformations are

particularly useful for segmentation in
which we want to isolate an object of
interest from a background as shown in
figure below
2.3 Intensity Transformation

S=log(1+r)………..(5)
We usually set c to 1. Grey levels must be in
the range [0.0, 1.0].
2.4 Logarithmic Transformations

The general form of the log transformation 2.5 Powers-Law Transformations

is The nth power and nth root curves shown in
fig. can be given by the expression,
s=c*log(1+r)………………………..(4)
s=crγ ……………………(6)
The log transformation maps a narrow range
of low input grey level values into a wider This transformation function is also called as
range of output values. The inverse log gamma correction. For various values of γ
transformation performs the opposite different levels of enhancements can be
transformation. Log functions are obtained. This technique is quite commonly
particularly useful when the input grey level called as Gamma Correction. If you notice,
values may have an extremely large range of different display monitors display images at
values. In the following example the Fourier different intensities and clarity. That means,
transform of an image is put through a log every monitor has built-in gamma correction
transform to reveal more detail in it with certain gamma ranges and so a
good monitor automatically corrects all the
images displayed on it for the best contrast
to give user the best experience. The
difference between the log transformations
Function and the power-law functions is that
using the power-law function a family of
possible transformation curves can be
obtained just by varying the λ. These are the
three basic image enhancement functions for
grey scale images that can be applied easily
Figure5. Example showing effect of
for any type of image for better contrast and
Logarithmic transformation
highlighting. Using the image negation
formula given above, it is not necessary for Figure 6. The images below show a contrast
the results to be mapped into the grey scale stretching linear transform to add contrast to a
range [0, L-1]. Output of L-1-r automatically poor quality image
falls in the range of [0, L-1]. But for the Log
and Power-Law transformations resulting
values are often quite distinctive, depending 2.7 Grey Level Slicing
upon control parameters like λ and
logarithmic scales. So the results of these Grey level slicing is the spatial domain
values should be mapped back to the grey equivalent to band-pass filtering. A grey
scale range to get a meaningful output level slicing function can either emphasize a
image. For example, Log function s = c log group of intensities and diminish all others
(1 + r) results in 0 and 2.41 for r varying or it can emphasize a group of grey levels
between 0 and 255, keeping c=1. So, the and leave the rest alone. Example is shown
range [0, 2.41] should be mapped to [0, L-1] in the following figure
for getting a meaningful image.

OUTPUT
Enter the value for c==>1
Enter the value for gamma==>1 % for
gamma value EQUALS TO 1

2.6 Piecewise Linear Transformation

Functions

Rather than using a well defined

mathematical function we can use arbitrary
user-defined transforms.

Figure 7. Showing example of Grey level

slicing

3. Histogram Processing

The histogram of a digital image with

intensity levels in the range [0, L-1] is a
discrete function
 Figure 8. The original image and its
histogram, and the equalized versions. Both
images are quantized to 64 grey levels

3.2 Histogram Matching

Histograms are frequently normalized by the
total number of pixels in the image. Histogram equalization automatically
Assuming an M x N image, a normalized determines a transformation function
histogram. seeking to produce an output image with
a uniform histogram. Another method is to
generate an image having a specified
histogram is histogram matching.
1. Find the histogram pr(r) of the input
image and determine its equalization
is related to probability of occurrence of rk in
transformation
the image.

………………
3.1 Histogram Equalization ……7

Histogram equalization is common 2. Use the specified pdf pz(r) of the output
technique for enhancing the appearance of image to obtain the transformation function:
images. Suppose we have an image which is
predominantly dark. Then its histogram
would be skewed towards the lower end of
the grey scale and all the image detail is ……………
compressed into the dark end of the …..8
histogram. If we could `stretch out' the grey
levels at the dark end to produce a more 3. Find the inverse transformation z = G-1(s)
uniformly distributed histogram then the – the mapping from s to z:
image would become much clearer.
…………………
…..9

4. Obtain the output image by equalizing the

input image first; then for each pixel in the
equalized image, perform the inverse
mapping to obtain the corresponding pixel
of the output image.
Histogram matching enables us to “match”
the grayscale distribution in one image to
the grayscale distribution in another image.
 Figure 10. (a) Original Image (b) Result of
global histogram Equalization (c) Result of
Local histogram equalization using 7x7
neighborhood about each pixel.

3.4 Use of Histogram Statistics for

Image Enhancement

Let the intensity in an image is represented

by a discrete rv r in [0, L-1] and let p (ri) is
the normalized histogram – estimate of pdf
Figure 9. Showing histogram Matching for the intensity. The nth statistical moment
different images is

3.3 Local Enhancement …………….1

0
Previous methods of histogram equalizations
and histogram matching are global. So, local For image intensities, a sample mean:
enhancement issued. Define square or
rectangular neighborhood (mask) and move
the center from pixel to pixel. For each ………………………..
neighborhood, calculate histogram of the 11
points in the neighborhood. Obtain
histogram equalization/specification and sample variance:
function. Map ray level of pixel centered in
neighborhood. It can use new pixel values
and previous histogram to calculate next ………………..1
histogram. 2

As previously, we may specify global mean

[11]-[12] and variance (for the entire image)
and local mean and variance for a specified
sub-image (subset of pixels).
 purpose contrast manipulation. For a dark
image, an expansion of gray levels is
accomplished using a power-law
transformation with a fractional exponent.
Log Transformation is Useful for enhancing
details in the darker regions of the image at
the expense of detail in the brighter regions
the higher-level values. For an image having
a washed-out appearance, a compression of
Figure 11. Showing example of using
gray levels is obtained using a power-law
histogram statistics for
transformation with γ
image enhancement
greater than 1. The histogram of an image
(i.e., a plot of the gray level frequencies)
provides important information regarding
Conclusion the contrast of an image. Histogram
Image enhancement algorithms offer a wide equalization is a transformation that
variety of approaches for modifying images stretches the contrast by redistributing the
to achieve visually acceptable images. The gray-level values uniformly. Only the global
choice of such techniques is a function of histogram equalization can be done
the specific task, image content, observer completely automatically Although we did
characteristics, and viewing conditions. The not discuss the computational cost of
point processing methods are most enhancement algorithms in this article it
primitive, yet essential image may play a critical role in choosing an
processing operations and are used primarily algorithm for real-time applications. Despite
for contrast enhancement. Image Negative is the effectiveness of each of these algorithms
suited for enhancing white detail embedded when applied separately, in practice one has
in dark regions and has applications in to devise a combination of such methods to
medical imaging. Power-law achieve more effective image enhancement.
transformations are useful for general

References
[1] Bhabatosh Chanda and Dwijest Dutta Majumder, 2002,Digital Image Processing and
Analysis.
[2] R.W.Jr. Weeks,(1996). Fundamental of Electronic Image Processing . Bellingham: SPIE
Press
[3] A. K. Jain, Fundamentals of Digital Image Processing. Englewood Cliffs, NJ: Prentice Hall,
1989.
[4] R.M. Haralick, and L.G. Shapiro, Computer and Robot Vision, Vol-1, Addison Wesley,
Reading, MA, 1992.
[5] R. Jain, R. Kasturi and B.G. Schunck, Machine Vision, McGraw-Hill International Edition,
1995.
[6] W. K. Pratt, Digital image processing, Prentice Hall, 1989.
[7] A.C. Bovik, Digital Image Processing Course Notes, Dept. of Electrical Engineering, U. of
Texas at Austin, 1995.
[8] J.C. Russ, The Image Processing Handbook, CRC Press, Boca Raton, FL., 1992.
[9] R Hummel, “Histogram modification techniques“, Computer Graphics and Image Processing,
Vol. 4, pp. 209-224, 1975.
[10] S. E. Umbaugh, “Computer Vision & Image Processing, “Prentice Hall PTR, 1998
[11] S. M. Pizer, et al., “Adaptive Histogram Equalization and its Variations,” Compute. Vision,
Graphics and Image Processing, Vol. 39, pp. 355-368, 1987.