'lg
RAJARATA UNIVERSITY OF SRI LANKA
FACULTY OF' APPLIED SCIENCES
B.Sc. (General) Degree in Applied Sciences
Second Year - Semester II Examination - November/ December 2016
COM 2304 _ COMPUTER GRAPHICS AND IMAGE PROCESSING
Time: Three (3) hours
Examination Index No:
Important Instru ction
.
.
.
.
.
.
s:
This paper has 4 questions in 14 pages.
Answer all questions (25 marks each).
Write your answers in English using the space provided in this question papor.
Do not tear off any part of this question paper.
Note that qurestions appear on both sides of the paper.
If a page is not printed, please inform the supen'isor immediately,
To be completed by the examiners:
Question numbers
Page 1 of 14
z$
Ouestion
1
Wnte short arlswers
a) Briefly explain hoi,v Artificial Intellisence
Conrputer Vision.
ard lmage Processinq are related to
(2 Marks)
b)
Why it is irnportant to study/ analyze about the Imase Source as an initial stage of Image
processir,g application development?
(3 Marks)
c)
What are the three basic quantities used to describe the properties of a chromatic light
source'/
(3 Marks)
d)
Justify the
following
statement using suitable exarnples;
"Depending on the nature of the sollrce, illumination energy is reflected from, or
transmitted through, obiects.
"
(4 Marks)
Page 2 of 14
30
e)
two techniques qsed to convert the continuous sensed data into digital form during
ttre digital image creation.
(2 Marks)
D
What is meant by Scan Conversion in computer graphic systerns?
(2 Marks)
g)
Compare CRT and LCD monitors using the given criteria
(3 Marks)
State
Cnteria
CRT
LCD
Size
Power Consumption
Image forming
h)
What is rneant by Imaqe Enhancement? Why it is problem
oriented?
(4 Marks)
Lnage Enhancement:
Why it is problem oriented?
i)
What is rneant by Image Interpolation?
Page 3 of 14
(2 Marks)
JT
Question 2
a)
List down five possible solrrces for irnage Noise.
(5 Marks)
b)
Explain the operation of following non linear filters
(4
Max Filter
c)
Maks)
Min Filter
Figure I depicts an image strip extracted along the horizontal direction of a digital image
given as fr, 1') where x and y are coordinate points. Each cell represents a single pixel of
tl.re image strip and the values indicate the intensities Apply following first order
derivative ftinction to the image strip to detect the edges.
l't order derivative of the image function
(/(x,
y))is
as
follows: {&
= .f
(*
+1)
- f (x)
Clearly state the values of the given irnage strip after applying the l't order derivative.
Explain the behavior of derivatives in the areas of constant intensity and intensity ramps.
(7 Marks)
+
1
l l
0
D
0
3
0
0
l
Figure 1: Image Strip
Page 4 of 74
3
I
)
0
c
c
7
7
32
d)
Write a shortnote on "@"
(5 Mtarks)
Page 5 of 14
JJ
e)
Figure 2 (a) and (b) depict two binary images. Explain the structural changes which can
occur in these images after applying the Erosion operations for Figure 2 (a) and Dilation
operation for Figure 2 (b).
(4 Marks)
Figure2-BinaryImages
Erosion for Figure 2 (a)
Dilation for Figure 2 (b)
Page 6 of 14
c1 ,
J't
O.uestion 3
a)
How nuny Kev Frames are required for a 4 nrinutes leugth cartoon video which has 20 fraures
per second? Note that it has been produced u'ith l0 in-between frames per key frarne. (2 }ularks)
b)
Explain l'row the following features will provide a lllore natural action for an animation. (2 Marks)
Straight ahead animation:
Staging:
c)
Apply Cohen-Sutherland Line Cliupins algorithn for clipping the following PrPz line. Cleary
shou' the algoritlm steps and iutersection points at clipping boundaries of the viewport Sketch
tlte clipping results in all iterations separately. Shaded area represents the viewport. (8 Ir{arks)
PageT of
t4
35
Iteration
0l
Iteration 02
I
Iteration 03
Iteration 04
I
y''
Iteration 05
Page 8 of 14
3S
Cohen-Sutherland Line Cli upine Aleorithm
:
Step 1 - Assign a region code for each endpoints,
Step 2 - Ifboth endpoints have a region code 0000 then accept this line.
$ep 3 - Else, perform the logical AND operation for both region codes.
Step 3. I - If the result is not 0000, then reject the line.
Step 3.2 - Else you need clipping,
Step 3.2.I - Choose an endpoint of the line that is outside the window.
Step 3,2.2 - Find the intersection point at lhe window boundary (base on region code).
Step 3.2.3 - Replace endpoint with the intersection point and update the region code.
Repeat step 2 until we find a clipped line either trivially accepted or trivially
Step 3.2.4
-
rejected.
Step 4 - Repeat step
d)
I for
other lines.
Apply Digital Differential Analyzer Alsorithm to draw line between (35, 12) a:nd (41,36)
points, Find only first ten coordinate points of this line. Finally, color the given grid cells to
visualize the line (origin is set as (35, 12) of this grid and is already rnarked in it), (8 Marks)
Step
x-
Y_
Plotted
Point
Plotted
Point
x
Rounded
value
35
t2
r
I
2
J
4
5
6
7
8
9
10
Page 9 of 14
---------___>
37
e)
Justify how to perfonrr following 2D transformation. "Rotate a polygon modeled about x
coordinates 3 and y coordinates 2 by 60u degrees and scale it by 1.5
1x
Polygon et {3, z} Point
gxpeded resuit
Page 10 of 14
tines."
(5 Marks)
38
Ouestion 04
a)
State the suitable OpenCV function names used to implement following Image
processing operations.
(3 Marks)
Description
Operation
To detect the edges of a digital rrnage based on the l" order
derivatives:
To detect the edges of a digital image based on hysteresis
technique:
To set a region of interest in a digital imase
To extract the video properties
:
:
To execute morphological gradrent operation:
To keep a RGB value:
b)
Comment on following OpenCV code segments.
(5 Marks)
cvMatchTemplate (im age,template,result, CV_TM_SQDIFF)
cvFilter2D(im age, im age, &kernel, cvPoint(- 1,-1 ));
cvSmooth (i m age,gaussian Im age,CV_GAUSSIAN,3,3)
Page 11 of 14
;
;
3$
c)
Figure 3 depicts a microscopic view of Paramecium cell specimen. Write an algorithm to
count the total number of cells in tliis specimen using image processing. Assume the
average area of a Parameciurn ceil is 780 pixels.
(6 Marks)
Figure 3 - Microscopic vierv of Paramecium Cells
Page 12 of 14
+CI
d)
Fill the blank code segments in the following C++ program which is written to count the
eyes of the cartoon shapes depicted in image.jpg (Figure 4, left image), Note that this
expected foreground can be extracted clearly using the binary threshold 29 (Figtve 4,
right image). Assume the given image.lpg is an 8 bit 3 channel image. Further, it has
bEen displayed the imagejpg on "Image" window and threshold result in "Threshold"
(11Marks)
window
Threshold; 29
al
* l**-
tt
tl
la
ta
Figure4-CartoonLnage
#include <highgui, h>;
#include <cv.h>1
#include <stdio. h>
using namespace std;
int g-thnesh =[lr
int main ( )
t
rprrmaget g-image =
cvloadlmar"<l--__-l
cvNamedl,'Jindow ( " Image ", 1) ;
cvShowlmage(
l-----l
,
l.;
g_image);
cvNamedWindow ( "Threshold Image",1);
Ipllmage* g_gnay = cvcneatelmage (
= cvcreateMemStorage (0).;
CvMemStorage* g_stonage
CvSeq* contours = 0;
cvcvtcolor (l
I
cvShowlmage( "'1-hnoshold lrrrago",
'l
[----l
Page 13 of 14
,8r 1) i
,GV_BGR2GRAy);
),
lt
ao
t,a
int
count =
_r.
cvFindConaor"rd-----l,
g-storage,
&contours, sizeof ( CvContour ),
cv_R ETR_L r ST, CV_CHAIN_Ap pROX_S rMp
cvPoint (0, 0) ) ;
cvWaitKey(
);
cvDestroyAl
Il,rlinciows ( ) ;
cvReleaselmrr"
I-----l
cvReleasermrr.,
L----l
neturn
l.;
f
,,
01
,,(r(t
*,1.*** END *i,r,**{. *
X(
Page 14 of 14
LE
,