DL - Assignment 11 Solution
DL - Assignment 11 Solution
Deep Learning
Assignment- Week 11
TYPE OF QUESTION: MCQ/MSQ
Number of questions: 10 Total mark: 10 X 1 = 10
______________________________________________________________________________
QUESTION 1:
What is the use of Skip Connection in image denoising networks?
Correct Answer: d
Detailed Solution:
QUESTION 2:
What will be the value of dice coefficient between A and B?
Detailed Solution:
|A|=7.82
|B|=8
A∩B = |A∩B|=7.42
QUESTION 3:
What will be the dice coefficient of following two one hot encoded vector? (|A|=no of 1 bit)
A 1 0 1 0 0 0 1 1 1 0 0 1 0 1
B 1 0 0 0 0 1 1 1 0 0 0 1 0 0
a. 0.83
b. 0.41
c. 0.67
d. 0.90
Correct Answer: c
Detailed Solution:
No of 1 bit in A=7
No of 1 bit in B =5
Overlapping 1 bit =4
2∗|𝐴∩𝐵| 2∗4
Dice Coefficient = |𝐴|+|𝐵|
= 5+7 = 0.67
______________________________________________________________________________
QUESTION 4:
Fully Connected Convolutional network or FCN became one of the major successful network
architectures. Can you identify what are the advantages of FCN which makes it a successful
architecture for semantic segmentation?
Correct Answer: d
Detailed Solution:
______________________________________________________________________________
QUESTION 5:
What are the different challenges one face while creating a facial recognition system?
Correct Answer: d
Detailed Solution:
Please refer to the lecture of week 11.
____________________________________________________________________________
QUESTION 6:
Suppose you have a 1𝐷 signal 𝑥 = [5,4,3,2,1] and a filter 𝑓 = [1,2,3,4], and you perform stride
2 transpose convolution on the signal 𝑥 by the filter 𝑓 to get the signal 𝑦. What will be the
signal 𝑦 if we don’t perform cropping?
a. 𝑦 = [1,2,5,8,9,14,13,20,19,26,3,4]
b. 𝑦 = [5, 10, 19, 28, 15, 22, 11, 16, 7, 10, 3, 4]
c. 𝑦 = [1,2,5,8,9,14,13,20,17,26,15,20]
d. 𝑦 = [0,0,5,8,9,14,13,20,19,26,0,0]
Correct Answer: b
Detailed Solution:
5 4 3 2 1
1*5=5 1
2*5=10 2
3*5+1*4=19 3 1
4*5+2*4=28 4 2
3*4+1*3=15 3 1
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
4*4+2*3=22 4 2
3*3+1*2=11 3 1
4*3+2*2=16 4 2
3*2+1*1=7 3 1
4*2+2*1=10 4 2
3*1=3 3
4*1=4 4
______________________________________________________________________________
____________________________________________________________________________
QUESTION 7:
Which of following can be a target output of semantic segmentation problem with 4 class?
a.
0 1 0 1 1 0 0 0 1 0 0 0
0 1 0 0 0 0 1 0 0 1 0 0
1 0 0 0 0 0 0 0 0 0 0 0
I II III IV
b.
0 1 0 1 0 0 0 0 1 0 1 0
0 1 0 0 1 0 1 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 1 1
I II III IV
c.
0 1 0 1 0 0 0 0 1 0 0 0
0 1 0 0 0 1 1 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 1 1
I II III IV
d.
0 1 0 1 0 0 0 0 1 0 0 0
0 1 0 1 0 0 1 0 0 0 0 0
1 1 0 0 1 0 0 0 1 0 1 1
I II III IV
Correct Answer: c
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
Detailed Solution:
Target output should be one hot encoded vector at every pixel location. It should one if the pixel
belongs to that particular class otherwise 0.
______________________________________________________________________________
QUESTION 8:
Which of the following is true for semantic segmentation?
Correct Answer: d
Detailed Solution:
____________________________________________________________________________
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
QUESTION 9:
In a Deep CNN architecture, the feature map before applying a max pool layer with (2x2) kernel
is given bellow.
50 30 3 4
14 12 7 10
44 1 14 19
2 5 16 20
After few successive convolution layers, the feature map is again up-sampled using Max Un-
pooling. If the following feature map is present before Max-Unpooling layer, what will be the
output of the Max-Unpooling layer?
50 10
44 20
a.
50 0 0 0
0 0 0 10
44 0 0 0
0 0 0 20
b.
50 0 10 0
0 0 0 0
44 0 20 0
0 0 0 0
c.
50 0 0 10
0 0 0 0
44 0 0 20
0 0 0 0
d.
0 0 0 0
0 50 0 10
0 0 0 0
0 44 0 20
NPTEL Online Certification Courses
Indian Institute of Technology Kharagpur
Correct Answer: a
Detailed Solution:
Please refer to lectures of week 11.
____________________________________________________________________________
QUESTION 10:
Which of the following operation reduces spatial dimension of features?
a. Max un-Pooling
b. Convolution with 3 × 3 Kernel, Stride=2, Padding all sides = 1
c. Convolution with 3 × 3 Kernel, Stride=1, Padding all sides = 1
d. Transposed convolution
Correct Answer: b
Detailed Solution:
____________________________________________________________________________
______________________________________________________________________________
************END*******