Questions tagged [fully-convolutional-networks]
For questions related to fully convolutional networks (FCNs), which is formally described in the paper "Fully Convolutional Networks for Semantic Segmentation" (2015) by Jonathan Long et al. An example of an FCN is the U-net (introduced in the paper "U-Net: Convolutional Networks for Biomedical Image Segmentation" by Olaf Ronneberger et al.).
18
questions
1
vote
1
answer
33
views
Why do we do need compression in Semantic Segmentation?
When doing semantic segmentation, we often make use of FCN, which can be thought of in two parts: an encoder and decoder. As I understand, the encoder compresses the image into a spatially small, but ...
- 113
0
votes
0
answers
20
views
How to compensate the receptive field offset in a Fully Convolutionl Network?
I'm studying the Fully Convolutional Networks right now and when it's clear that the receptive field is not dependent on the input size (the whole network in a way is independent from the input size), ...
- 1
0
votes
1
answer
71
views
Can a fully convolutional network handle smaller images at test time?
It is said that a fully convolutional network can handle any image size. I don't understand.
Unlike regular CNN, a fully convolutional network reinterprets the dense layer as a convolutional operation....
- 111
0
votes
0
answers
66
views
Position-wise FC layer vs standard fully-connected layer
What exactly is the position-wise FC layer in the transformer?
I understand the FC layer, bascially, each unit in the next layer is a linear combination of the each other input. I don't know how does ...
- 183
0
votes
1
answer
180
views
Why FCNN is not always better than CNN?
Why Fully-Connected Neural Network is not always better than Convolutional Neural Network?
FCNN is easily overfitting due to many params, then why didn't it reduce the params to reduce overfitting.
If ...
2
votes
1
answer
273
views
Best practice for handling letterboxed images for non fully-convolutional deep learning networks?
I'm working on a depth estimation network. It has two outputs:
A relative depth map
A scalar for scaling the relative depth map into an absolute depth map. This second output uses dense layers so ...
- 153
1
vote
0
answers
86
views
FCNs: Questions about the filter rarefaction in the CVPR paper [Long et al., 2015]
I am reading the paper about the fully convolutional network (FCN).
I had some questions about the part where the authors discuss the filter rarefaction technique (I guess this is roughly equivalent ...
- 11
0
votes
0
answers
476
views
Computational complexity of a CNN network
In the following network, the convolution operations of convolutional blocks are performed by three 1-D kernels with the sizes 8, 5, and 3 respectively along with stride equal to 1. The final network ...
- 1
1
vote
0
answers
22
views
In the DeepView paper, do they use the same FCN for all depth slices AND all views?
I'm trying to replicate a paper from Google on view synthesis/lightfields from 2019: DeepView: View Synthesis with Learned Gradient Descent and this is the PDF.
Basically the input to the neural ...
- 11
19
votes
1
answer
32k
views
What is a fully convolution network?
I was surveying some literature related to Fully Convolutional Networks and came across the following phrase,
A fully convolutional network is achieved by replacing the parameter-rich fully ...
- 325
8
votes
2
answers
727
views
What is the point of using 1D and 2D convolutions with a kernel size of 1 and 1x1 respectively?
I understand the gist of what convolutional neural networks do and what they are used for, but I still wrestle a bit with how they function on a conceptual level. For example, I get that filters with ...
- 81
2
votes
1
answer
325
views
Can a fully convolutional network always return an image of the same size as the original?
I'm trying to perform a segmentation task on images of multiple sizes using fully convolutional neural networks.
Currently, I'm using EfficientNet as a feature extractor, and adding a deconvolution/...
- 137
3
votes
3
answers
584
views
Does a fully convolutional network share the same translation invariance properties we get from networks that use max-pooling?
Does a fully convolutional network share the same translation invariance properties we get from networks that use max-pooling?
If not, why do they perform as well as networks which use max-pooling?
- 793
4
votes
1
answer
3k
views
What do the words "coarse" and "fine" mean in the context of computer vision?
I was reading the well know paper Fully Convolutional Networks for Semantic Segmentation, and, throughout the whole paper, they talk use the term fine and coarse. I was wondering what they mean. The ...
- 211
6
votes
1
answer
3k
views
Why can a fully convolutional network accept images of any size?
On this article, it says that:
The UNET was developed by Olaf Ronneberger et al. for Bio Medical Image Segmentation. The architecture contains two paths. First path is the contraction path (also ...
- 163
10
votes
2
answers
15k
views
How to handle rectangular images in convolutional neural networks?
Almost all the convolutional neural network architecture I have come across have a square input size of an image, like $32 \times 32$, $64 \times 64$ or $128 \times 128$. Ideally, we might not have a ...
33
votes
5
answers
41k
views
How can I deal with images of variable dimensions when doing image segmentation?
I'm facing the problem of having images of different dimensions as inputs in a segmentation task. Note that the images do not even have the same aspect ratio.
One common approach that I found in ...
- 597
2
votes
1
answer
877
views
What are the counterparts of non-linearities and dropout in fully convolutional networks?
I am trying to replicate the fully convolutional networks (FCN) concept described here for semantic segmentation. It seems people have successfully trained such models by removing fully connected ...
- 129