Image Enhancement Based on
Histogram
1 April 26, 2024
Overview
Image enhancement using point processing
techniques
Negative Transformation
Log transformation
Power law Transformation
Contrast Stretching
Gray level Slicing
Bit Plane slicing.
2 April 26, 2024
This topic covers
What is histogram?
Histogram equalization
Histogram
Specification/Matching/Modification
3 April 26, 2024
Definition
Histogram of an image represents the relative frequency of
occurrence of the various gray levels in an image in range [0, L-1].
Example:
Gray Number of The histogram is a discrete function
level pixels h(rk) h(rk) = nk
0 40
40 – Where rk is the kth gray level
1 20
– nk is no. of pixels in image
2 10 having gray level rk
20
3 15
15
10 10
4 10 3 2
5 3 0 1 2 3 4 5 6
Gray levels
4 6 2 April 26, 2024
OR Normalized Histogram
Histogram of an image represents Probability occurrences of the
various gray levels in an Image.
Example: • The normalised histogram is given by p(rk) =
Gray Number of Probability nk/n for k=0,1,2,….,L-1
level pixels (nk) (nk/n)
p(rk)
0 40 0.4
0.4
1 20 0.2
2 10 0.1
3 15 0.15 0.2
4 10 0.1 0.15
0.1 0.1
5 3 0.03 0.03 0.02
6 2 0.02 0 1 2 3 4 5 6
Gray levels
5
Total April 26, 2024
n=100
Example
6 April 26, 2024
Components of histogram are concentrated on
the (low) dark side of gray scale
Components of histogram are concentrated on
the high (Light) side of gray scale
Low contrast image has narrow histogram and
centered in the middle of gray scale
The high contrast image has the most evenly spaced
histogram.
7 April 26, 2024
Uniform Histogram
We can conclude…
An image whose pixels tend to occupy the entire range
of possible gray levels and tend to be distributed
uniformly, will have an appearance of high contrast
and will display a large variety of gray tones.
Histogram Transformation Uniform/flat
of input Function histogram
image
Histogram Equalisation
8 April 26, 2024
Histogram Equalisation
We need to define transformation that transforms
bad histogram in flat (uniform) histogram.
We know s=T(r) where 0≤ r ≤L-1
But what should be T????
9 April 26, 2024
Histogram Equalisation
The transformation T must satisfy 2 conditions
s T (r ) 0 r L 1
a. T(r) is a monotonically increasing function in the interval 0 r L -1;
(i.eT (r2 ) T (r1 ) for r2 r1 )
b. 0 T (r ) L -1 for 0 r L -1.
10 April 26, 2024
Histogram Equalisation
The transformation T must satisfy 2 conditions
s T (r ) 0 r L 1
a. T(r) is a monotonically increasing function in the interval 0 r L -1;
(i.eT (r2 ) T (r1 ) for r2 r1 )
b. 0 T (r ) L -1 for 0 r L -1.
Condition a) guarantees that output intensity values will
never be less than corresponding input intensities values.
b) Guarantees that range of output intensities is same as the
input.
11 April 26, 2024
Histogram Equalization
k nj
L 1 k
( L 1) nj k=0,1,..., L-1
j 0 MN MN j 0
April 26, 2024 12
Histogram Equalisation (cont..)
k
sk T (rk ) ( L 1) pr (r j )
j 0
Let k=0 to 7
s0=(L-1)pr(r0) This mapping is called
s1=(L-1)[pr (r0)+pr (r1)] Histogram Equalisation
s2=(L-1 )[pr (r0)+pr (r1)+pr (r2)]
.
.
s7=(L-1)[pr (r0)+pr (r1)+pr (r2)+…..+pr (r7)]
The output image (s) is obtained by mapping each
pixel with level rk in input image into
corresponding
13
pixel with level sk in output image
April 26, 2024
Example
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
Histogram of original image
h(rk) 1023
850
790 656
329
245 122
81
0 1 2 3 4 5 6 7
Gray levels
14 April 26, 2024
k
sk T (rk ) ( L 1) pr (rj )
Example j 0
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
Grey Nk
Level
0 790
1 1023
2 850
3 656
4 329
5 245
6 122
7 81
Total=4096
15 April 26, 2024
k
sk T (rk ) ( L 1) pr (rj )
Example j 0
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
Grey Nk Pr(rk)=nk/n
Level
0 790 0.19
1 1023 0.25
2 850 0.21
3 656 0.16
4 329 0.08
5 245 0.06
6 122 0.03
7 81 0.02
Total=4096
16 April 26, 2024
k
sk T (rk ) ( L 1) pr (rj )
Example j 0
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
k
Grey Nk Pr(rk)=nk/n
Level p (r )
j 0
r j
0 790 0.19 0.19
1 1023 0.25 0.44
2 850 0.21 0.65
3 656 0.16 0.81
4 329 0.08 0.89
5 245 0.06 0.95
6 122 0.03 0.98
7 81 0.02 1
Total=4096
17 April 26, 2024
k
sk T (rk ) ( L 1) pr (rj )
Example j 0
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
k k
Grey Nk Pr(rk)=nk/n
Level p (r )
j 0
r j ( L 1) pr (rj )
j 0
0 790 0.19 0.19 1.33
1 1023 0.25 0.44 3.08
2 850 0.21 0.65 4.55
3 656 0.16 0.81 5.67
4 329 0.08 0.89 6.23
5 245 0.06 0.95 6.65
6 122 0.03 0.98 6.86
7 81 0.02 1 7
Total=4096
18 April 26, 2024
k
sk T (rk ) ( L 1) pr (rj )
Example j 0
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
Grey Nk Pr(rk)=nk/n s k p (r ) ( L 1) k p (r ) new gray
Level k j 0
r j
j 0
r j
level
0 790 0.19 0.19 1.33 1
1 1023 0.25 0.44 3.08 3
2 850 0.21 0.65 4.55 5
3 656 0.16 0.81 5.67 6
4 329 0.08 0.89 6.23 6
5 245 0.06 0.95 6.65 7
6 122 0.03 0.98 6.86 7
7 81 0.02 1 7 7
Total=4096
19 April 26, 2024
Old Gray Nk new gray New Gray Level Nk
Level level
1 790
0 790 1
3 1023
1 1023 3
5 850
2 850 5
6 656+329=985
3 656 6
7 245+122+81=448
4 329 6
5 245 7
6 122 7
7 81 7
Total=4096
20 April 26, 2024
New Gray Level Nk
0 0
1 790
2 0
3 1023
4 0
5 850
6 656+329=985
7 245+122+81=448
21 April 26, 2024
h(rk) 1023
850
790
Histogram of original Image
656
329
245
122
81
0 1 2 3 4 5 6 7
Gray levels
h(rk) 1023 985
850
790
Histogram of equalized Image
448
22 April 26, 2024
0 1 2 3 4 5 6 7
Gray levels
Equalisation Transformation Function
23 April 26, 2024
24 April 26, 2024
Original Image Equalised Image
Histogram of Original Image Histogram of Equalised Image
2000 2000
1500 1500
1000 1000
500 500
0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
25 April 26, 2024
26 April 26, 2024
Histogram equalization
Perform histogram equalization and draw a new
equalized histogram of the following image data
Gray Level 0 1 2 3 4 5 6 7
No. of 550 900 650 150 300 250 110 90
pixels
27 April 26, 2024
Example
0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1
1 1 1 2 2 2 3 4 Perform histogram equalization of
1 1 1 2 2 2 3 4 the given image.
1 1 1 2 2 2 3 4
1 1 1 2 2 2 3 5
2 2 2 2 2 5 5 5
6 6 6 6 7 7 5 5
28 April 26, 2024
Example
Old Gray Nk new gray
Level level
0 0 0 0 0 0 0 0
0 8 1
1 1 1 1 1 1 1 1
1 20 3
1 1 1 2 2 2 3 4
2 17 5
1 1 1 2 2 2 3 4
3 4 5
1 1 1 2 2 2 3 4
1 1 1 2 2 2 3 5 4 3 6
2 2 2 2 2 5 5 5 5 6 6
6 6 6 6 7 7 5 5 6 4 7
7 2 7
Total=64
29 April 26, 2024
New Gray Level Nk
0 0
1 8
2 0
3 20
4 0
5 17+4=21
6 3+6=9
7 4+2=6
30 April 26, 2024
Histogram equalisation Cont..
31 April 26, 2024
Example cont….
32 April 26, 2024
Histogram Equalisation (Cont….)
For Some applications histogram equalization
is not the best approach.
It is useful sometimes to be able to specify the
shape of histogram that we wish the
processed image to have.
The method to generate a processed image
that has specified histogram is called
Histogram matching or Histogram
specification
33 April 26, 2024
Histogram Specification
The transformation function T(rk) for the given histogram
Pr(rj) , j=0, 1 ,2, ….L-1 is given by
k
sk T (rk ) pr (rj )
j 0
k nj
( L 1)
j 0 MN
The transformation function G(z q) for the specified
histogram Pz(zi) , i=0, 1 ,2, ….L-1 is given by
q q
ni
G ( z q ) p z ( zi ) ( L 1) sk
34 i 0 i 0 MN April 26, 2024
Histogram Specification
So we are seeking the value of z that satisfy
the equation
sk T (rk )
G ( zk ) sk zk G 1 ( sk )
If don’t able to find exact G-1 , find closest z
35 April 26, 2024
Steps for histogram specification
Obtain the histogram of original image.
Equalize the histogram of original image (Obtain T(r) i. e s)
Equalize the histogram of specified image. (Obtain G(z))
Perform the mapping process so that the values of step 2 can
be mapped to the result of step 3. (i.e. Obtain G-1 )
36 April 26, 2024
Example of histogram specification
Given histogram (a) and (b). Modify (a) as given by histogram (b)
Histogram (a)
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 790 1023 850 656 329 245 122 81
Histogram (b)
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 0 0 0 614 819 1230 819 614
37 April 26, 2024
s=T(r) G(z)
Old Gray nk new gray Old Gray new gray
Level rk level sk Level zk level G(zk)
0 790 1
1 1023 3
0 0 G(zk)=sk
1 0
2 850 5 zk = G-1(s)
2 0
3 6
985 3 1
4 6
5 7 4 2
6 448 7 5 5
7 7 6 6
Total=4096 7 7
38 April 26, 2024
T(r) G(z) Final Result z= G-1(s)
Old Gray Nk new gray Gray Level new gray G-1(1)=3
Level level (Sk) zk level G(zk)
G-1(3)=4
0 790 1 G-1(5)=5
0 0 G-1(6)=6
1 1023 3
1 0 G-1(7)=7
2 850 5
2 0
3 6 Gray Nk
985 3 1 Level
4 6
4 2 0 0
5 7
5 5 1 0
6 448 7
6 6 2 0
7 7
7 7 3 790
Total=4096
4 1023
5 850
6 985
7 448
Total=4096
39 April 26, 2024
Specified Histogram
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 0 0 0 614 819 1230 819 614
Output Histogram
Gray Level 0 1 2 3 4 5 6 7
No. of pixels 0 0 0 790 1023 850 985 448
40 April 26, 2024
Original Image
41 April 26, 2024
Histogram Equalisation
42 April 26, 2024
(a)Specified Histogram
Enhanced image after applying Histogram
Specification
43 Histogram of above image April 26, 2024
