19
votes
Accepted
When should I use 3D convolutions?
3D convolutions are used when you want to extract features in 3 dimensions or establish a relationship between 3 dimensions.
Essentially, it's the same as 2D convolutions, but the kernel movement is ...
- 1,419
12
votes
Accepted
How can the convolution operation be implemented as a matrix multiplication?
To show how the convolution (in the context of CNNs) can be viewed as matrix-vector multiplication, let's suppose that we want to apply a $3 \times 3$ kernel to a $4 \times 4$ input, with no padding ...
- 41.4k
9
votes
Accepted
Do convolutional neural networks perform convolution or cross-correlation?
Short answer
Theoretically, convolutional neural networks (CNNs) can either perform the cross-correlation or convolution: it does not really matter whether they perform the cross-correlation or ...
- 41.4k
6
votes
When should I use 3D convolutions?
3D convolutions should be used when you want to extract spatial features from your input on 3 dimensions. For computer vision, they are typically used on volumetric images, which are 3D.
Some examples ...
- 201
6
votes
What is the difference between graph convolution in the spatial vs spectral domain?
Spectral Convolution
In a spectral graph convolution, we perform an Eigen decomposition of the Laplacian Matrix of the graph. This Eigen decomposition helps us in understanding the underlying ...
6
votes
Is there anything that ensures that convolutional filters don't end up the same?
No, nothing really prevents the weights from being different. In practice though they end up almost always different because it makes the model more expressive (i.e. more powerful), so gradient ...
- 1,690
6
votes
Accepted
What is a filter in the context of graph convolutional networks?
Short answer
Check out the paper of Shuman et al. [1], it provides some background on Graph Signal Processing, including answers to your questions in sections II.C and III.A
Long Answer
Question 1
Yes,...
5
votes
What does 'input planes' mean in the phrase 'input signal/image composed of several input planes'?
Yes, it is a bit misleading. What it really means is input channels, so it would be: nn.Conv2d: Applies a 2D convolution over an input signal composed of several input channels.
So, why don't just use ...
- 1,168
4
votes
Why would you implement the position-wise feed-forward network of the transformer with convolution layers?
I'm going to post another guess to this question - it won't be a complete answer, but hopefully it'll provide some direction towards finding a more legitimate answer.
The feed-forward networks as ...
- 91
4
votes
Does each filter in each convolution layer create a new image?
For a 3 channel image (RGB), each filter in a convolutional layer computes a feature map which is essentially a single channel image. Typically, 2D convolutional filters are used for multichannel ...
- 156
4
votes
Accepted
Does each filter in each convolution layer create a new image?
You are partially correct. On CNNs the output shape per layer is defined by the amount of filters used, and the application of the filters (dilation, stride, padding, etc.).
CNNs shapes
In your ...
- 156
4
votes
Does each filter in each convolution layer create a new image?
About the images inside the CNN layers: I really recommend this article since there is no one short answer to this question and it probably will be better to experiment with it.
About the RGB input ...
- 96
4
votes
Accepted
How is the convolution layer is usually implemented in practice?
I don't think that to understand convolution you need to dig into the nested code of huge libraries, since the code becomes quickly really hard to understand and convoluted (ba dum tsss!). Joking ...
- 5,428
4
votes
Accepted
When should we use separable convolution?
Context of the question
This is a link to the text cited in the question.
It refers to the usage of SeparableConv2D (tf, keras name). A related question on StackOverflow is "What is the ...
- 1,293
4
votes
Are there any advantages of the local attention against convolutions?
It is true that when using local attention with a window of size 5, the "receptive field" is the same as a CNN with kernel size 5 (or two CNN layers with kernel size 3). However, there is a ...
- 161
3
votes
Accepted
Can I shuffle image channel data as a form of data augmentation?
As a rule of thumb for image data augmentation, look at the augmented images:
Can you correctly classify or measure your target label from the augmented images?
Could something similar to the ...
- 33.3k
3
votes
Can CNNs be applied to non-image data, given that the convolution and pooling operations are mainly applied to imagery?
You can use CNN for time-series data. The Convolutional Recurrent Neural Network (RCNN) is one of the examples.
Convolutional layers basically extract features from images. It is not related to time-...
- 71
3
votes
Accepted
Can CNNs be applied to non-image data, given that the convolution and pooling operations are mainly applied to imagery?
Usually, you need to ensure that your convolutions are causal, meaning that there is no information leakage from the future into the past. You could start by looking at this paper, which compares ...
- 1,158
3
votes
Accepted
How do I optimize the number of filters in a convolution layer?
Am I right in thinking that because there are only newImageX * newImageY patterns in the 32 x 32 image, that the maximum amount ...
- 41.4k
3
votes
Why would you implement the position-wise feed-forward network of the transformer with convolution layers?
1) The math is the exact same, so from an optimization or mathematical perspective there is no difference
2) Here are my guesses to a possible answer.
Habit: People may just call one over the ...
- 2,387
3
votes
Wouldn't convolutional neural network models work better without flattening the input in any stages?
I have had similar thoughts about neural networks before. Convolution layers are layers of two dimensional nodes effectively passing the spacial data so why don't we use two dimensional hidden layers ...
- 555
3
votes
Accepted
What is the difference between asymmetric and depthwise separable convolution?
They are not the same thing.
asymmetric convolutions work by taking the x and y axes of the image separately. For example performing a convolution with an $(n \times 1)$ kernel before one with a $(...
- 2,387
3
votes
Accepted
In which scenario would you want to have two adjacent pooling layers?
Assuming you're not referring to any particular type of pooling operation, it's possible that you could have, for example, a mean pool followed by a max or min pool. What this could do is combine the ...
- 527
3
votes
Accepted
How is the depth of the input related to the depth of the output of a convolutional layer?
The reason why you go from 16 to 3 channels is that, in a 2d convolution, filters span the entire depth of the input. Therefore, your filters would actually be $7 \times 7 \times 16$ in order to cover ...
- 953
3
votes
What is the use of the regular convolutional layer in expansion path of U-Net?
The point is that in the expansive path you have two forms of information:
the information from the contracting path, which includes all high-level features extracted from the original image.
the ...
- 3,213
3
votes
What is the difference between same convolution and full convolution in terms of feature map size?
What happens to the size of output feature map in case of full convolution?
It increases.
First one is valid padding: the blue square is not padded, so the green square is smaller. Third one is ...
- 179
3
votes
Accepted
How is the convolution operation connected to neural networks?
The convolution operation performed by most CNNs that you will find (on the web) assumes that the signals/functions are discrete and 2-dimensional (e.g. images can be viewed as 2-dimensional discrete ...
- 41.4k
3
votes
Accepted
Is down-sampling the only purpose of using stride?
The general purpose of stride (along with padding) is to determine the spatial dimensions of the output. So, with appropriate stride (and padding), you can also make the spatial dimensions of the ...
- 41.4k
3
votes
Accepted
What's the best criterion for evaluating activation maps in a CNN?
It looks like you are thinking about something like UNET architecture.
The final layer to your toy problem with a "dot" on the nose can be modelled as a convolution layer from last ...
- 498
Only top scored, non community-wiki answers of a minimum length are eligible