Deep Learning A-Z™: Convolutional Neural Networks (CNN) - Step 2: Pooling

•

2 likes•1,694 views

Kirill Eremenko

Education

© SuperDataScienceDeep Learning A-Z
Image Source: Wikipedia

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
Max Pooling
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
1 1
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
Max Pooling
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
1 1 0
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4 2
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4 2 1
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4 2 1
0
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4 2 1
0 2
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Feature Map
0 1 0 0 0
0 1 1 1 0
1 0 1 2 1
1 4 2 1 0
0 0 1 2 1
Max Pooling
1 1 0
4 2 1
0 2 1
Pooled Feature Map

© SuperDataScienceDeep Learning A-Z
Evaluation of Pooling Operations in
Convolutional Architectures for
Object Recognition
By Dominik Scherer et al. (2010)
Link:
http://ais.uni-bonn.de/papers/icann2010_maxpool.pdf
Additional Reading:

© SuperDataScienceDeep Learning A-Z
Input Image
Convolutional Layer
0 0 0 0 0 0 0
0 1 0 0 0 1 0
0 0 0 0 0 0 0
0 0 0 1 0 0 0
0 1 0 0 0 1 0
0 0 1 1 1 0 0
0 0 0 0 0 0 0
Pooling Layer
Convolution Pooling

© SuperDataScienceDeep Learning A-Z
Image Source: scs.ryerson.ca/~aharley/vis/conv/flat.html

What's hot

Deep Learning A-Z™: AutoEncoders - AutoEncodersKirill Eremenko

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - Module 3Kirill Eremenko

Deep Learning A-Z™: Artificial Neural Networks (ANN) - Plan of AttackKirill Eremenko

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - LSTMsKirill Eremenko

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - LSTM VariationsKirill Eremenko

Deep Learning A-Z™: Self Organizing Maps (SOM) - How do SOMs learn (part 2)Kirill Eremenko

Deep Learning A-Z™: Artificial Neural Networks (ANN) - BackpropagationKirill Eremenko

Deep Learning A-Z™: Boltzmann Machine - Module 5 - Boltzmann MachineKirill Eremenko

Deep Learning A-Z™: Artificial Neural Networks (ANN) - The NeuronKirill Eremenko

Deep Learning A-Z™: Self Organizing Maps (SOM) - Reading an Advanced SOMKirill Eremenko

Deep Learning A-Z™: Artificial Neural Networks (ANN) - Gradient DescentKirill Eremenko

Deep Learning A-Z™: Boltzmann Machines - Module 5Kirill Eremenko

Deep Learning A-Z™: Convolutional Neural Networks (CNN) - SummaryKirill Eremenko

Deep Learning A-Z™: Self Organizing Maps (SOM) - How Do SOMs Work?Kirill Eremenko

Deep Learning A-Z™: Self Organizing Maps (SOM) - Module 4Kirill Eremenko

Deep Learning A-Z™: Boltzmann Machines - Restricted Boltzmann MachineKirill Eremenko

Deep Learning A-Z™: Boltzmann Machines - Energy Based Models (ebm)Kirill Eremenko

Deep Learning A-Z™: AutoEncoders - Training an AutoEncoderKirill Eremenko

Deep Learning A-Z™: Self Organizing Maps (SOM) - Plan of attackKirill Eremenko

Deep Learning A-Z™: Boltzmann Machines - Boltzmann MachineKirill Eremenko

What's hot (20)

Deep Learning A-Z™: AutoEncoders - AutoEncoders

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - Module 3

Deep Learning A-Z™: Artificial Neural Networks (ANN) - Plan of Attack

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - LSTMs

Deep Learning A-Z™: Recurrent Neural Networks (RNN) - LSTM Variations

Deep Learning A-Z™: Self Organizing Maps (SOM) - How do SOMs learn (part 2)

Deep Learning A-Z™: Artificial Neural Networks (ANN) - Backpropagation

Deep Learning A-Z™: Boltzmann Machine - Module 5 - Boltzmann Machine

Deep Learning A-Z™: Artificial Neural Networks (ANN) - The Neuron

Deep Learning A-Z™: Self Organizing Maps (SOM) - Reading an Advanced SOM

Deep Learning A-Z™: Artificial Neural Networks (ANN) - Gradient Descent

Deep Learning A-Z™: Boltzmann Machines - Module 5

Deep Learning A-Z™: Convolutional Neural Networks (CNN) - Summary

Deep Learning A-Z™: Self Organizing Maps (SOM) - How Do SOMs Work?

Deep Learning A-Z™: Self Organizing Maps (SOM) - Module 4

Deep Learning A-Z™: Boltzmann Machines - Restricted Boltzmann Machine

Deep Learning A-Z™: Boltzmann Machines - Energy Based Models (ebm)

Deep Learning A-Z™: AutoEncoders - Training an AutoEncoder

Deep Learning A-Z™: Self Organizing Maps (SOM) - Plan of attack

Deep Learning A-Z™: Boltzmann Machines - Boltzmann Machine

Recently uploaded

CELL CYCLE Division Science 8 quarter IV.pptxJiesonDelaCerna

CARE OF CHILD IN INCUBATOR..........pptxGaneshChakor2

POINT- BIOCHEMISTRY SEM 2 ENZYMES UNIT 5.pptxSayali Powar

TataKelola dan KamSiber Kecerdasan Buatan v022.pdfSarwono Sutikno, Dr.Eng.,CISA,CISSP,CISM,CSX-F

Biting mechanism of poisonous snakes.pdfadityarao40181

Framing an Appropriate Research Question 6b9b26d93da94caf993c038d9efcdedb.pdfUjwalaBharambe

Full Stack Web Development Course for BeginnersSabitha Banu

Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Celine George

How to Make a Pirate ship Primary Education.pptxmanuelaromero2013

“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...Marc Dusseiller Dusjagr

Pharmacognosy Flower 3. Compositae 2023.pdfMahmoud M. Sallam

EPANDING THE CONTENT OF AN OUTLINE using notes.pptxRaymartEstabillo3

DATA STRUCTURE AND ALGORITHM for beginnersSabitha Banu

Crayon Activity Handout For the Crayon AUnboundStockton

Historical philosophical, theoretical, and legal foundations of special and i...jaredbarbolino94

Proudly South Africa powerpoint Thorisha.pptxthorishapillay1

Interactive Powerpoint_How to Master effective communicationnomboosow

Final demo Grade 9 for demo Plan dessert.pptxAvyJaneVismanos

Capitol Tech U Doctoral Presentation - April 2024.pptxCapitolTechU

9953330565 Low Rate Call Girls In Rohini Delhi NCR9953056974 Low Rate Call Girls In Saket, Delhi NCR

Recently uploaded (20)

CELL CYCLE Division Science 8 quarter IV.pptx

CARE OF CHILD IN INCUBATOR..........pptx

POINT- BIOCHEMISTRY SEM 2 ENZYMES UNIT 5.pptx

TataKelola dan KamSiber Kecerdasan Buatan v022.pdf

Biting mechanism of poisonous snakes.pdf

Framing an Appropriate Research Question 6b9b26d93da94caf993c038d9efcdedb.pdf

Full Stack Web Development Course for Beginners

Incoming and Outgoing Shipments in 1 STEP Using Odoo 17

How to Make a Pirate ship Primary Education.pptx

“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...

Pharmacognosy Flower 3. Compositae 2023.pdf

EPANDING THE CONTENT OF AN OUTLINE using notes.pptx

DATA STRUCTURE AND ALGORITHM for beginners

Crayon Activity Handout For the Crayon A

Historical philosophical, theoretical, and legal foundations of special and i...

Proudly South Africa powerpoint Thorisha.pptx

Interactive Powerpoint_How to Master effective communication

Final demo Grade 9 for demo Plan dessert.pptx

Capitol Tech U Doctoral Presentation - April 2024.pptx

9953330565 Low Rate Call Girls In Rohini Delhi NCR

Deep Learning A-Z™: Convolutional Neural Networks (CNN) - Step 2: Pooling

5. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling Pooled Feature Map

6. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 Pooled Feature Map

7. © SuperDataScienceDeep Learning A-Z Feature Map Max Pooling 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 1 1 Pooled Feature Map

8. © SuperDataScienceDeep Learning A-Z Feature Map Max Pooling 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 1 1 0 Pooled Feature Map

9. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 Pooled Feature Map

10. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 2 Pooled Feature Map

11. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 2 1 Pooled Feature Map

12. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 2 1 0 Pooled Feature Map

13. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 2 1 0 2 Pooled Feature Map

14. © SuperDataScienceDeep Learning A-Z Feature Map 0 1 0 0 0 0 1 1 1 0 1 0 1 2 1 1 4 2 1 0 0 0 1 2 1 Max Pooling 1 1 0 4 2 1 0 2 1 Pooled Feature Map

15. © SuperDataScienceDeep Learning A-Z Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition By Dominik Scherer et al. (2010) Link: http://ais.uni-bonn.de/papers/icann2010_maxpool.pdf Additional Reading:

16. © SuperDataScienceDeep Learning A-Z Input Image Convolutional Layer 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 Pooling Layer Convolution Pooling

Deep Learning A-Z™: Convolutional Neural Networks (CNN) - Step 2: Pooling

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

More from Kirill Eremenko

More from Kirill Eremenko (17)

Recently uploaded

Recently uploaded (20)

Deep Learning A-Z™: Convolutional Neural Networks (CNN) - Step 2: Pooling