Name : ……………………………………………………………
htt
Roll No. : ……………………………………………..…………..
Invigilator’s Signature : ………………………………………..
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
p:/
2010
DIGITAL IMAGE PROCESSING
Time Allotted : 3 Hours Full Marks : 70
/q
The figures in the margin indicate full marks.
Candidates are required to give their answers in their own words
pap
as far as practicable.
GROUP – A
( Multiple Choice Type Questions )
er.
1. Choose the correct alternatives for any ten of the
following : 10 × 1 = 10
i) Which one of the following transform coding systems
wb
( usually ) does not decompose the input image into
several sub-images before transform ?
a) Discrete Fourier transform coding
b) Walse-Hadamard transform coding
ut .
c) Discrete Cosine Transform coding
d) Wavelet Transform coding.
ii) Huffman coding approach reduces
a c.
a) coding redundancy only
b) inter-pixel redundancy only
c) coding & inter-pixel redundancy
d) psycho-visual redundancy only.
in
8216 [ Turn over
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
iii) To obtain the impulse response of filter the input
htt
impulse image should be like
a) a total white image of size M × N
b) a total black image of size M × N
p:/
c) a white dot in a centre of black image of M × N
d) a black dot in a centre of white image of M × N.
/q
iv) Which one of the following coding approaches attacks
both the coding and inter-pixel redundancy ?
pap
a) Huffman coding b) LZW coding
c) B2 coding d) All of these.
v) The relation of intensity ( I ) and R, G & B in RGB colour
! model is
er.
a) I = 0·6R + 0·25G + 0·15B
(R + G + B)
b) I=
3
wb
(R + 2 G + B)
c) I=
4
!
d) I = 0·5R + 0·25G + 0·25B.
ut .
vi)! If the image is degraded by motion blur and added noise
then .................... gives the best result
a) median filter
a c.
b) inverse filter
c) Wiener filter
d) constraint least square filter.
in
8216 2
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
vii) Diagonal edge can be detected by using which of the
htt
following masks ?
a) 0 1 0
1 –4 1
p:/
0 1 0
b) 1 1 1
/q
1 1 1
1 1 1
pap
c) 0 1 2
–1 0 1
–2 –1 0
er.
d) –1 –1 –1
–1 8 –1
wb
–1 –1 –1
viii) Faulty switching introduces
a) Gaussian noise
ut .
b) Rayleigh noise
c) Gamma noise
d) Impulse noise.
a c.
ix) Poor illumination introduces
a) Gaussian noise b) Rayleigh noise
c) Exponential noise d) Impulse noise.
in
8216 3 [ Turn over
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
x) Which of the following grey level transformations
produces image negative ?
htt
a) S = C log ( 1 + r )
b) S=L–1–r
p:/
c) S = C rk
nj
d) S k = "kj = 0 , k = 0, 1, 2, 3, ... ( L # 1) .
n
!
/q
xi) Erosion
! a) sharps a region b) blurs a region
pap
c) increases a region d) decreases a region.
xii) Euclidian distance of two points ( x, y ) and ( s, t ) of a
two-dimensional space is
1
2 2 2
a) [( x " s) + ( y " t) ]
er.
b) |x – s| + |y – t|
! c) Max (|x – s|, |y – t|)
wb
d) none of these.
GROUP – B
( Short Answer Type Questions )
ut .
Answer any three of the following. 3 × 5 = 15
2. Prove that imaginary part of a Fourier transform of an even
function is zero.
a c.
3. Show that the Fourier transform of the auto-correlation
2
function of f ( x ) is the power spectrum " (u) .
in
8216 4
!
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
4. Suppose a digital image is subjected to histogram
htt
equalization. Show that a second pass of histogram will
produce exactly the same result as the first pass ?
p:/
5. Discuss the limiting effect of repeatedly applying a 3 × 3
spatial filter to a digital image. Ignore the border effects.
6. Develop a procedure for computing the median of an n × n
/q
neighbourhood. Propose a technique for updating the median
as the centre of neighbourhood if moved from pixel to pixel.
pap
7. What is pixel ? Explain 4-neighbour and 8-neighbour of a
pixel. Explain m-adjacency. 1+2+2
er.
8. Explain image sensing and acquisition ( using single sensor,
sensor strip and sensor arrays ).
wb
GROUP – C
( Long Answer Type Questions )
Answer any three of the following. 3 × 15 = 45
ut .
9. a) What effect would, setting to zero the lower order bit
planes, have on the histogram of an image, in general ?
a c.
b) What would be observed on the histogram if higher
order bit planes are set to ‘0’ ?
c) Obtain the Haar Transform matrix for N = 8. 3+3+9
in
8216 5 [ Turn over
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
10. a) Under what condition the does Butterworth low-pass
htt
filter become an ideal low-pass filter ?
b) Show that a high-pass filtered image in frequency
p:/
domain can be obtained by using the method of
subtracting a low-pass filtered image from the original.
/q
c) An image is blurred by uniform acceleration in x
direction. If the image is at rest at time t = 0 and
at2
pap
accelerate with an acceleration X 0 (t) = for a time T.
2
Find the transfer function H ( u, v ) to reconstruct the
image. ! 3+7+5
er.
11. a) Can variable length coding procedure be used to
compress a histogram equalized image with two gray
levels ?
wb
b) Can such an image contain inter-pixel redundancies
that could be exploited for data compression ?
ut .
c) Find a set of code words and word length using
Huffman coding scheme for a set of input gray levels
a c.
with probabilities as given below :
Symbol : S1 S2 S3 S4 S5 S6 S7 S8
Probability : 0·02 0·15 0·03 0·15 0·05 0·2 0·1 0·3
in
8216 6
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
d) Image blurring caused by long term exposure to
htt
atmospheric turbulence can be modelled by the transfer
function H (u, v) =
((
exp " u2 + v2 )) . Assume negligible
2
2#
p:/
noise.
!
What is the equation of Wiener filter to be used to
/q
reconstruct an image blurred by this type of
pap
degradation ? 3+2+5+5
12. a) Construct the entire 4 bit gray code.
b) Create a general procedure for converting a gray code
er.
number to its equivalent binary and use it to decode
0111010100111.
wb
c) The arithmetic decoding process is the reverse of the
encoding process. Decode the message 0·23355 using
ut .
the given coding model. Consider ‘!’ as the terminating
symbol.
a c.
Symbol : A E I O U !
Probability : 0·2 0·3 0·1 0·2 0·1 0·1
4+3+8
in
8216 7 [ Turn over
CS/[Link] (ECE-NEW)/SEM-8/EC-803D/2010
13. a) Briefly describe any three colour models.
htt
b) Write the conversion rules for converting RGB colour
model to HSI colour model and from HSI to RGB.
p:/
c) How a colour image can be converted into a gray scale
image ? 6+6+3
/q
14. a) Explain the chain code.
pap
b) What is boundary descriptor ?
c) Explain Fourier descriptor. 8+2+5
15. a) Describe LZW coding with example.
er.
b) What is transform coding ? Give some examples of
transform coding. 10 + 5
wb
ut .
a c.
in
8216 8
