Question 1 (30 marks)

a) The following figures, A and B demonstrate the effects of two intensity

transformations. Analyze the different types of transformation these figures fit to in.
Briefly explain how each of the transformation is different from the other.
(6 marks)

Figure (A) Figure (B)

Figure A – Contrast Stretching, Figure B – Thresholding

You need to briefly explain each transformation

b) The following figures denote two input matrix (A) and mask (B) generated using a
simple transformation(s) applied to the images. Determine the output matrix for the
highlighted pixel. Write down the intermediate steps, if there are any.
(4 marks)

Mask (B)

Input Matrix (A)

Answer

c) "Histogram equalization derives the transformation function automatically from

probability density function." Assess whether the statement is correct or not. Justify
your answer.
(2 marks)
Yes. The statement is correct.
Transformed using its cumulative probability distribution

1
d) The following figure (A) denotes histogram of an image in spatial domain. Briefly
explain the effect on the image and suggest the transformation function that fulfills
the figure.
(4 marks)

Figure (A)

Identity transforms function.

You need to explain the effect on an image if you apply Identity transforms
function.

e) The histograms of input image Figure (A) and a processed image Figure (B) are given
below. Determine the quality of the input image and suggest the most suitable
technique which could be used to satisfy the visual appearance given in figure (B).

(4 marks)

Figure (A)
Figure (B)
Answer:

Low contrast image, to adjust most suitable technique is histogram equalization.

2
f) The input image is a 3-bit image of size 20x20 pixels (MN = 400). It has the intensity
distribution shown in the following table, where the intensity levels are integers in the
range [0, L-1] = [0, 7]. Apply the histogram equalization transformation function
using the probability density function P(rk) values for different values of input
intensities rk .

Copy the table given below to the answer booklet and carry out the histogram
equalization for the part of the input image given below.
(10 marks)

r nk p(rk) ∑ P(𝑟𝑘 ) s

0 70
1 100
2 40
3 60
4 0
5 80
6 10
7 40

Answer

r nk p(rk) ∑ P(𝑟𝑘 ) s
0 70 70/400 = 0.175 0.175 7*0.175 = 1
1 100 100/400 = 0.25 0.175+0.25=0.425 7*0.425 = 3
2 40 40/400 = 0.1 0.425+0.1=0.525 7*0.525 = 4
3 60 60/400 = 0.15 0.525+0.15=0.675 7*0.675 = 5

4 0 0/400 = 0.0 0.675+0.0=0.675 7*0.675 = 5

5 80 80/400 = 0.2 0.675+0.2=0.875 7*0.875 = 6

6 10 10/400 = 0.025 0.875+0.025=0.900 7*0.900 = 6

7 40 40/400 = 0.1 0.900+0.1=1.0 7*1.0 =7

3
Question 2 (25 Marks)
a) You were asked to construct an edge detector using a Laplacian filter based on a
second order derivative function, briefly explain the second order derivative
responses at a ramp edge and flat intensity area using an example.
(6 marks)

Second Order Derivative Filter:

• Zero in flat areas

• Non zero at onset/start and end of a gray level step or ramp
• Zero along ramps of constant slope

Take an example and calculate the responses.

b) The following figure shows part of an 8-bit image (A). Apply the filter (B) given
below to calculate the filter response of the bounded section of the input image (A)
given below. You may need to handle boundary pixels using replication.
(6 marks)

1/9 1/9 1/9

1/9 1/9 1/9
1/9 1/9 1/9
Filter (B)

Input Image Matrix (A)

Answer
=4*1/9+4*1/9+4*1/9+4*1/9+4*1/9+4*1/9+4*1/9+4*1/9+3*1/9
=4

4
c) The following figure (A) indicates two regions separated by a vertical edge. Plot the
grey level profile along the vertical line of 1D image. Comment on the first and
second derivative response near the edge.
(5 marks)

Figure (A)

Answer

Comment on the first and second derivative response near the edge.

1st order response 2nd order response

a) Intensity metrics of a small part of an 8 bit gray scale image (A) is given below.
Calculate the output pixel values after applying filter (B) only for the highlighted
area. Show your calculations as well. You may need to handle boundary pixels using
replication technique.
(6 marks)

Filter (B)

Input Image Matrix (A)

5
 Answer

b) "Ramp slope is inversely proportional to the degree of blurring”. Assess whether the
statement is correct or not. Justify your answer.

(2 marks)
Statement is correct.

Explain with a diagram.

6
Question 3 (30 Marks)
a) “The basic effect of the dilation operator on a binary image is to gradually shrink the
boundaries of regions of foreground pixels”. Do you agree with this statement?
Assess your answer.
(4 marks)

No the statement is incorrect.

Refer the basic effect of the operator on a binary image and explain here.

b) Briefly explain the morphological operation(s) for shape detection using set
operators.
(4 marks)

Hit-or-Miss Transform
Explain with a example.

c) The following figure shows the result (B) of applying mathematical morphology
operation(s) to the original image (A). Critique the technique(s) used to obtain the
output (B).
(4 marks)

Input Image (A) Output Image (B)

Answer: Morphological Opening

7
d) The following figure (F1) demonstrates an application of object detection. Briefly
explain the technique(s) used to obtain the intermediate outputs from an input image
(A).
(4 marks)

Figure F1

Image B - Binary image

Image C- region filled image
Image D – Boundary extraction

e) Briefly explain the basic effects of the following process with respect to background
and foreground pixels.
(4 marks)

Kernel (B)

Input Image (A)

8
 Answer

f) If the shaded area represents the foreground pixels, find the results of Closing operation
performed on the image given in Figure (A), using the structuring elements shown in
Figure (B). Note that the origin is highlighted. Show your answers in the given
worksheet.
(10 marks)

Origin

1 1 1
1 1 1
1 1 1
Fig. (B)

Fig. (A)

Answer

9
Question 4 (15 Marks)
a) “Color is a very important attribute in image processing with lots of extra
information”. Do you agree with this statement? Assess your answer.
(3 marks)
Yes. Agree.
Color is a powerful descriptor which can be used for various tasks: simplify
image analysis, segmentation, object detection, extraction and processing based
on color, tracking and identification.

The human visual system can distinguish hundreds of thousands of different

color shades and intensities, but only around 100 shades of grey.

Therefore, Color is a very important attribute in image processing with lots of

extra information.

b) Color models provide a standard way to specify a particular color, by defining a 3D

coordinate system, and a subspace that contains all constructible colors within a
particular model. Briefly explain one additive color model.
(3 marks)

RGB- Additive model

 The RGB model is used for color monitors and most video cameras.
 The colors present in the light add to form new colors, and is appropriate for the
mixing of colored light for example.
 It can be viewed as 3 monochrome intensity images (red/green/blue).
 This is an additive model, it asks what is added to black to get a particular color.

10
c)
(i). The following Chromaticity diagram shows all visible colors to human vision. If
you draw straight line segment joining two points, Point E, Point 540, what will be
defined by that line? Briefly explain the importance of that line.
(3 marks)

Straight line segment joining two points define all colors that can be created by
mixing these two colors additively.

(ii). Briefly explain where pure spectrum colors are located using the diagram.
(3 marks)

Pure spectrum colors are located around the boundary

(iii). Refer to the above diagram explain the point where and how standard light is
produced.
(3 marks)

Point E

Point of equal energy corresponds to equal fractions of the three primary colors,
red, green and blue. This is considered as CIE standard for white light.

11