Lecture # 13:

Image Enhancement:-
Principle Objective of Enhancement

 To Process an image so that the result will be more suitable

than the original image for a specific application.

 The suitableness is up to each application.

 A method which is quite useful for enhancing an image may

not necessarily be the best approach for enhancing another
images

2
2 Domains

 Spatial Domain : (image plane)

 Techniques are based on direct manipulation of pixels in an
image

 Frequency Domain :
 Techniques are based on modifying the Fourier transform
of an image

 There are some enhancement techniques based on various

combinations of methods from these two categories.

3
Good Images
 For human visualization

 The visual evaluation of image quality is a highly subjective

process.
 It is hard to standardize the definition of a good image.

 For machine perception

 The evaluation task is easier.

 A good image is one which gives the best machine
recognition results.

 A certain amount of trial and error usually is required before a

particular image enhancement approach is selected.

4
Spatial Domain

 It includes Procedures that

operate directly on pixels.

g(x,y) = T[f(x,y)]

where:
 f(x,y) is the input image

 g(x,y) is the processed image

 T is an operator on f defined
over some neighborhood of
(x,y)

5
Mask /Filter

 Neighborhood of a point (x,y) can

be defined by using a square/
(x,y) rectangular (common used) or
circular subimage area centered at
(x,y)
•
 The center of the subimage is
moved from pixel to pixel starting
at the top of the corner

6
Point Processing

 Neighborhood = 1x1 pixel

 g depends on only the value of f at (x,y)
 T = gray level (or intensity or mapping) transformation
function s = T(r)

 Where
 r = gray level of f(x,y)
 s = gray level of g(x,y)

7
Contrast Stretching

 Produces higher contrast than

the original by:

 Darkening the levels

below m in the original
image

 Brightening the levels

above m in the original
image

8
Contrast Stretching

 A limiting Case, where the

possible outputs includes 0 &
1 only.

 Produce a two-level (binary)

image

9
Mask Processing /Filtering

 Neighborhood is larger than 1x1 pixel.

 Use a function of the values of input Image f in a predefined

neighborhood of (x,y) to determine the value of output Image
g at (x,y)

 The value of the mask coefficients determine the nature of the

process.

 Used in techniques
 Image Sharpening
 Image Smoothing

10
 Basic Gray Level Transformations

 Linear function
Negative  Negative
nth root  Identity transformations
Output gray level, s

Log
nth power
 Logarithm function
 Log
 Inverse-log transformation

 Power-law function
Identity Inverse Log  nth power
 nth root transformations

Input gray level, r

11
 IDENTITY Function

Negative

nth root
 Output intensities are
identical to input intensities.
Output gray level, s

Log
nth power

 Is included in the graph only

for completeness.

 The output Image is identical

Identity Inverse Log to the Input Image.

Input gray level, r

12
 NEGATIVE Function

 An image with gray level in the

Negative
range [0, L-1] can have

nth root
 Negative transformation :
s = L – 1 –r
Output gray level, s

Log
nth power
 Reversing the intensity levels of
an image.

 Suitable for enhancing white or

gray detail embedded in dark
Identity Inverse Log regions of an image, especially
when the black area dominant in
size.
Input gray level, r
13
 LOG Function

Negative
s = c log (1+r)
nth root
c is a constant
r0
Log
Output gray level, s

nth power
 Log curve maps a narrow
range of low gray-level values
in the input image into a
wider range of output levels.

Identity Inverse Log  Used to expand the values of

dark pixels in an image while
compressing the higher-level
values.
Input gray level, r

14
LOG Function

 It compresses the dynamic range of images with large

variations in pixel values.

 Example of image with dynamic range: Fourier spectrum image

 It can have intensity range from 0 to 106 or higher.

 We can’t see the significant degree of detail as it will be lost in

the display.

15
Example of Logarithm Image

Fourier Spectrum with range = 0 to Result after apply the log

1.5 x 106 transformation with c = 1, range =
0 to 6.2

16
Inverse Logarithm Transformations

 Used to expand the values of high pixels in an image while

compressing the darker-level values.

17
 Inverse Logarithm Transformations

s = cr
 c and  are positive constants
Output gray level, s

 Power-law curves with

fractional values of  map a
narrow range of dark input
values into a wider range of
output values.

 If c =  = 1, It becomes an
Identity function
Input gray level, r
Plots of s = cr for various values of 
(c = 1 in all cases)

Used to expand the values of high pixels in an image while compressing the 18
darker-level values.
Gaama Correction

Monitor  Cathode ray tube (CRT)

devices have an intensity-
to-voltage response that is
a power function, with 
varying from 1.8 to 2.5
Gamma
 = 2.5
correction

 The picture will become

darker.

Monitor  Gamma correction is done

by preprocessing the image
before inputting it to the
monitor with s = cr1/
 =1/2.5 = 0.4
19
 a b
Another Example MRI c d

(a) A magnetic resonance image of an

upper thoracic human spine with a
fracture dislocation and spinal cord
impingement
 The picture is predominately dark
 An expansion of gray levels are
desirable  needs  < 1
(b) result after power-law
transformation with  = 0.6, c=1
(c) transformation with  = 0.4
(best result)
(d) transformation with  = 0.3
(under acceptable level)

20
Effect of decreasing gamma

 When the  is reduced too much, the image

begins to reduce contrast to the point where the
image started to have very slight “wash-out” look,
especially in the background

21
 a b
Another Example c d

(a) image has a washed-out

appearance, it needs a
compression of gray levels 
needs  > 1
(b) result after power-law
transformation with  = 3.0
(suitable)
(c) transformation with  = 4.0
(suitable)
(d) transformation with  = 5.0
(high contrast, the image has
areas that are too dark, some
detail is lost)

22
Contrast Stretching
 (a)increase the dynamic range
of the gray levels in the image

 (b) a low-contrast image :

result from poor illumination,
lack of dynamic range in the
imaging sensor

 (c) result of contrast stretching:

(r1,s1) = (rmin,0) and (r2,s2) = (rmax,L-1)

 (d) result of thresholding

23
Gray Level Slicing

 Highlighting a specific range

of gray levels in an image
 Display a high value of all
gray levels in the range of
interest and a low value
for all other gray levels

 (a) transformation highlights

range [A,B] of gray level and
reduces all others to a
constant level

 (b) transformation highlights

range [A,B] but preserves all
other levels

a b 24
 Bit-plane slicing

 Highlighting the contribution

made to total image appearance
One 8-bit byte
Bit-plane 7 by specific bits.
(most significant)

 Suppose each pixel is

represented by 8 bits.

 Higher-order bits contain the

majority of the visually
significant data.
Bit-plane 0
(least significant)
 Useful for analyzing the relative
importance played by each bit of
the image.

25
Example

 The (binary) image for bit-

plane 7 can be obtained by
processing the input image
with a thresholding gray-level
transformation.

 Map all levels between 0

and 127 to 0
 Map all levels between
129 and 255 to 255

An 8-bit fractal image

26
8 Bit Planes

Bit-plane Bit-plane
7 6

Bit- Bit- Bit-

plane plane plane
5 4 3
Bit- Bit- Bit-
plane plane plane
2 1 0

27