25/05/2018 Deep Learning For Visual Computing - - Unit 7 - Week 6

reviewer1@nptel.iitm.ac.in ▼

Courses » Deep Learning For Visual Computing

Announcements Course Ask a Question Progress Mentor

Unit 7 - Week 6

Course
outline Assignment 6
The due date for submitting this assignment has passed. Due on 2018-03-07, 23:59 IST.
How to access
the portal Submitted assignment

Week 1 1) What should be the response dimension if an image of dimension 3 × 25 × 25 is 1 point

convoloved with a filter of dimension 3 × 5 × 5 ?
Week 2

1 × 21 × 21
Week 3

Week 4 3 × 25 × 25

Week 5 1 × 25 × 25

Week 6 25 × 25 × 1

Lecture 26: No, the answer is incorrect.

Convolutional Score: 0
Neural Network
Building Blocks Accepted Answers:
1 × 21 × 21
Lecture 27:
Simple CNN 2) What should be the response dimension if an image of dimension 3 × 25 × 25 is padded 1 point
Model: LeNet
with width of 2 and convoloved with a filter of dimension 3 × 5 × 5 ?
Lecture 28:
LeNet Definition
1 × 21 × 21
Lecture 29:
Training a
LeNet for 3 × 25 × 25

MNIST
Classification
1 × 25 × 25

Lecture 30:
Modifying a 25 × 25 × 1
LeNet for
CIFAR No, the answer is incorrect.
Feedback for
Score: 0
Week 6 Accepted Answers:
1 × 25 × 25
Week 6: Lecture
Slides
3) What should be the response dimension if an image of dimension 3 × 25 × 25 is 1 point
Quiz : convoloved with a filter of dimension 3 × 5 × 5 and considered stride as 2 ?
Assignment 6

Week 6:
1 × 11 × 11
Assignment
Solutions
3 × 12.5 × 12.5
Week 7
1 × 12.5 × 12.5

https://onlinecourses.nptel.ac.in/noc18_ee08/unit?unit=10&assessment=94 1/4
25/05/2018 Deep Learning For Visual Computing - - Unit 7 - Week 6

Week 8
1 × 12 × 12

Week 9 No, the answer is incorrect.

Score: 0
Week 10
Accepted Answers:
1 × 11 × 11
Week 11
4) Including maxpooling equips the model to be invariant to with in the 1 point
⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
Week 12 limits of maxpooling kernel

DOWNLOAD
VIDEOS Scale

Translation

Rotation

None of the above

No, the answer is incorrect.

Score: 0
Accepted Answers:
Translation

5) Which of the following filter is functionally equivalent to maxpooling operation 1 point

Erosion
Dilation
Rank Filter
None of the above

No, the answer is incorrect.

Score: 0
Accepted Answers:
Dilation

6) If the encoder part of an autoencoder with 784 input neurons and 100 hidden neurons can 1 point
be rewrapped as a n convolutional filters each with a, b and c as input channels, filter height and filter
width. What are the values of n, a, b and c

Hint: Autoencoder is trained on gray scale image patches of size 28 × 28

1, 100, 28 and 28
100, 1, 28 and 28
28, 100, 1 and 28
None of the abpve

No, the answer is incorrect.

Score: 0
Accepted Answers:
100, 1, 28 and 28

7) Which operation needs max pooling indices or switch variables? 1 point

Pooling
Unpooling
Activation
Deconvolution

No, the answer is incorrect.

Score: 0
Accepted Answers:
Unpooling

https://onlinecourses.nptel.ac.in/noc18_ee08/unit?unit=10&assessment=94 2/4
25/05/2018 Deep Learning For Visual Computing - - Unit 7 - Week 6
8) Which operation needs learnable weights? 1 point

Pooling
Unpooling
Activation
Deconvolution

No, the answer is incorrect.

Score: 0
Accepted Answers:
Deconvolution

9) What of the following activation function has no gradient for negative responses 1 point

Sigmoid
Tanh
ReLU
Absolute

No, the answer is incorrect.

Score: 0
Accepted Answers:
ReLU

10)LeNet is designed for classification of 1 point

Alphabets
Numbers or digits
Images
None of the above

No, the answer is incorrect.

Score: 0
Accepted Answers:
Numbers or digits

Previous Page End

© 2014 NPTEL - Privacy & Terms - Honor Code - FAQs -

A project of In association with

Funded by

Powered by

https://onlinecourses.nptel.ac.in/noc18_ee08/unit?unit=10&assessment=94 3/4
25/05/2018 Deep Learning For Visual Computing - - Unit 7 - Week 6

https://onlinecourses.nptel.ac.in/noc18_ee08/unit?unit=10&assessment=94 4/4