2017-11-15
Feature map for 1st image (X) :
| 2 | 0 | 1 | 0 |
| 0 | 2 | 0 | 1 |
| 1 | 0 | 2 | 0 |
| 0 | 1 | 0 | 2 |
Feature Map for 2nd Image (+):
| 0 | 1 | 1 | 0 |
| 1 | 1 | 2 | 1 |
| 1 | 2 | 1 | 1 |
| 0 | 1 | 1 | 0 |
4x4 = 16 neurons
(Edited: 2017-11-15) Feature map for 1st image (X) :
{|
|-
| 2 || 0 || 1 || 0
|-
| 0 || 2 || 0 || 1
|-
| 1 || 0 || 2 || 0
|-
| 0 || 1 || 0 || 2
|}
Feature Map for 2nd Image (+):
{|
|-
| 0 || 1 || 1 || 0
|-
| 1 || 1 || 2 || 1
|-
| 1 || 2 || 1 || 1
|-
| 0 || 1 || 1 || 0
|}
4x4 = 16 neurons
((resource:WhatsApp Image 2017-11-15 at 5.23.24 PM.jpeg|Resource Description for WhatsApp Image 2017-11-15 at 5.23.24 PM.jpeg))
Name: Abhinav Tipirisetty
((resource:WhatsApp Image 2017-11-15 at 5.22.42 PM.jpeg|Resource Description for WhatsApp Image 2017-11-15 at 5.22.42 PM.jpeg))
Name: Abhinav Tipirisetty
Using a threshold function(with value 2) for feature map1:
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Using a threshold function(with value 2) for feature map2:
0 0 0 0
0 0 1 0
0 1 0 0
0 0 0 0
Number of neurons = 16
(Edited: 2017-11-15) Using a threshold function(with value 2) for feature map1:
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Using a threshold function(with value 2) for feature map2:
0 0 0 0
0 0 1 0
0 1 0 0
0 0 0 0
Number of neurons = 16
the first image: after using the kernel as a filter
[[1 0 0 0],[0 1 0 0],[0 0 1 0 ],[0 0 0 1]]
the first image: after using the kernel as a filter
[[1 0 0 0],[0 1 0 0],[0 0 1 0 ],[0 0 0 1]]
Result of first image
[[2,0,1,0],[0,2,0,1],[1,0,2,0],[0,1,0,2]]
Result of second Image:
[[0,1,1,0],[1,1,2,1],[1,2,1,1],[0,1,1,0]]
Result of first image
[[2,0,1,0],[0,2,0,1],[1,0,2,0],[0,1,0,2]]
Result of second Image:
[[0,1,1,0],[1,1,2,1],[1,2,1,1],[0,1,1,0]]
((resource:Photo Nov 15, 17 49 44.jpg|Resource Description for Photo Nov 15, 17 49 44.jpg))
We are using the identity function as our activation.
Our filter is given by
` [[1,0], [0,1]] `
Input for first neuron is
`[[1,0], [0,1]]`
So, output of neuron 1 of first image is dot product of this input with filter, equal to 2.
Input for second neuron is
` [[0,0], [1,0]]`
output of neuron 2 of first image is dot product of this input with filter, equal to 0.
Doing like this, for each of the neuron, we get, the feature maps as:
` f1 = [[2,0,1,0], [0,2,0,1], [1,0,2,0], [0,1,0,2]]`
` f2 = [[0,1,1,0], [1,1,2,1], [1,2,1,1], [0,1,1,0]]`
There are 16 neurons in each feature map.
(Edited: 2017-11-15) We are using the identity function as our activation.
Our filter is given by
@BT@ [[1,0], [0,1]] @BT@
Input for first neuron is
@BT@[[1,0], [0,1]]@BT@
So, output of neuron 1 of first image is dot product of this input with filter, equal to 2.
Input for second neuron is
@BT@ [[0,0], [1,0]]@BT@
output of neuron 2 of first image is dot product of this input with filter, equal to 0.
Doing like this, for each of the neuron, we get, the feature maps as:
@BT@ f1 = [[2,0,1,0], [0,2,0,1], [1,0,2,0], [0,1,0,2]]@BT@
@BT@ f2 = [[0,1,1,0], [1,1,2,1], [1,2,1,1], [0,1,1,0]]@BT@
There are 16 neurons in each feature map.
((resource:In-class Exercise.jpeg|Resource Description for In-class Exercise.jpeg))
((resource:IMG_4883.jpg|Resource Description for IMG_4883.jpg))
(c) 2024 Yioop - PHP Search Engine