NPTEL Online Certification Courses

Indian Institute of Technology Kharagpur

Digital Image Processing

Assignment- Week 10
TYPE OF QUESTION: MCQ/MSQ
Number of questions: 10 Total mark: 10 X 1 = 10
______________________________________________________________________________

QUESTION 1:

Which of the following statement(s) is (are) true?

Statement 1: The drawback of Wiener filter is that constant approximation of a ratio of power
spectra is not always suitable
Statement 2: In constraint least square filtering using Iterative Selection of Gamma requires
knowledge of mean and variance of noise

a) Statement 1 is True; Statement 2 is False

b) Statement 1 is True; Statement 2 is True

c) Statement 1 is False; Statement 2 is True

d) Statement 1 is False; Statement 2 is False

Correct Answer: b.

Detailed Solution:

In Wiener filter is that constant approximation of a ratio of power spectra is not always suitable.

Constraint least square filtering requires noise parameter estimation for iterative search of
Gamma.

Hence option b is correct.

_______________________________________________________________
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

QUESTION 2:
Which of the following the most suitable metric for match or mismatch measure in template
matching?
a) Sum of squared difference.

b) Cross co-relation

c) Normalised cross-correlation

d) None of these

Correct Answer: c.

Detailed Solution:

Normalized cross correlation is the most suitable metric for match or mismatch measure in
template matching as it normalizes the correlation with the matching image window pixels.

Hence option c is correct.

______________________________________________________________________________

QUESTION 3:
In which of the following application(s), image registration is used.
a) Template Matching

b) Mosaicing

c) Image Fusion

d) All of these

Correct Answer: d.
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

Detailed Solution:

Template matching, mosaicking, image fusion are all application of image registration as
these requires some sort of matching and alignment between images.

Hence option d is correct.

______________________________________________________________________________

QUESTION 4:

In the minimum error function given by the expression

 
 1 H (u , v
2 
Fˆ (u , v) =  G (u , v)
 H (u , v) H (u , v 2 + S n (u , v) 
 S f (u , v) 

Where, 𝑆𝑆𝑓𝑓 (𝑢𝑢, 𝑣𝑣) is Power Spectrum of Original Image, 𝑆𝑆𝑛𝑛 (𝑢𝑢, 𝑣𝑣) is Noise Power
Spectrum, 𝐻𝐻(𝑢𝑢, 𝑣𝑣) is degradation function. The Wiener Filter is equal to the Inverse filter when

a) 𝑆𝑆𝑛𝑛 (𝑢𝑢, 𝑣𝑣) = 1

b) 𝑆𝑆𝑛𝑛 (𝑢𝑢, 𝑣𝑣) = 0

c) 𝑆𝑆𝑛𝑛 (𝑢𝑢, 𝑣𝑣) = 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 (except 0, 1)

d) None of the above

Correct Answer: b.

Detailed Solution:

If we substitute 𝑆𝑆𝑛𝑛 (𝑢𝑢, 𝑣𝑣) = 0 in the following relation

 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

 
2
∧  1 H ( u, v ) 
F ( u, v ) =  .  G ( u , v ) , then it becomes
 H ( u , v ) H ( u , v ) 2 + Sη (u , v) 

 Sf ( u , v ) 
∧  1 
F ( u, v ) =   G ( u , v ) , which is nothing but inverse filtering.
 H ( u, v ) 
Hence option b is correct.
______________________________________________________________________________

QUESTION 5:

In which of the following restoration techniques, noise parameters estimation is required?

a) Constrained least square filtering

b) Weiner filtering

c) Inverse filtering

d) All of these

Correct Answer: a.

Detailed Solution:

Noise parameter estimation is required for image restoration using constrained least square
filtering.

Hence option a is correct.

______________________________________________________________________________
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

QUESTION 6:
In Minimum Mean Square Error Filtering, which of the following is/are true?

I. Both the degradation function and statistical characteristics of noise are considered for
restoration.

II. Both the image and noise are considered as random processes and objective is to find an
estimate.

a) Only I
b) Only II
c) Both I and II
d) None of the above mentioned.

Correct Answer: c.

Detailed Solution:

In Minimum Mean Square Error Filtering,

Both the degradation function and statistical characteristics of noise are considered for
restoration.

Both the image and noise are considered as random processes and objective is to find an
estimate.

Hence option c is correct.

_
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

_____________________________________________________________________________

QUESTION 7:
The advantage of the Wiener filter over Inverse Filter is due to which of the following options?

a) Transfer function of Wiener Filter is chosen to minimise the mean square error
using statistical information of only image fields.
b) Transfer function of Wiener Filter is chosen to minimise the mean square error
using statistical information of only noise fields.
c) Transfer function of Wiener Filter is chosen to minimise the mean square error
using statistical information of both image and noise fields.
d) None of the above

Correct Answer: c.

Detailed Solution:

Transfer function of Wiener Filter is chosen to minimise the mean square error using statistical
information of both image and noise fields.

Hence option c is correct.

______________________________________________________________________________

QUESTION 8:

For image registration of two images, minimum point correspondences are required to establish
pixel relations of an object?

a) 8

b) 2

c) 6

d) 4
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

Correct Answer: d.

Detailed Solution:

Let (x', y') be pixel coordinates in one image and (x, y) pixel coordinates in other image. These are
related as x' = k1x+k2y+k3xy+k4 and y' = k4x+k6y+k7xy+k8.

We have 8 unknowns and each correspondence pair gives us two equations. So we require at least
4 point correspondence.

Hence option d is correct.

______________________________________________________________________________

QUESTION 9:

Which of the following filter can be used for remove noise from image given below? (Noisy
image given in the left side and original noise free image is given in the right side.)

a) Low pass filter

b) Band pass filter
c) High pass filter
d) Band reject filter

Correct Answer: d.
Detailed Solution:
 NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur

The given image is corrupted via periodic noise. In case of periodic noise, we have
significantly higher response at frequencies equal to the frequency of the periodic noise and
these can be eliminated with the help of a band reject filter.

______________________________________________________________________________

QUESTION 10:
Assume the degradation model for atmospheric turbulence is defined in frequency domain as
2 2
H (u , v) = e − K (u +v )5 / 6 . If the value of K is decreased, then atmospheric turbulence __________.

a) Decreases

b) Remains constant

c) Increases

d) None of these

Correct Answer: a.

Detailed Solution:

If the value of K is decreased, then atmospheric turbulence also decreases.

If K= 0 H ( u,v ) = 1 , that means input and output is same and no degradation

Hence option a is correct.

______________________________________________________________________________

************END*******