The Stack Overflow podcast is back! Listen to an interview with our new CEO.

Questions tagged [convolutional-neural-networks]

For questions about convolutional neural networks, also known as CNN or ConvNet.

Filter by
Sorted by
Tagged with
2
votes
1answer
77 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 (...
2
votes
1answer
49 views

What is the purpose and benefit of applying CNN to a graph?

I'm new to the graph convolution network. I wonder what is the main purpose of applying data with graph structure to CNN?
1
vote
0answers
29 views

Keras CRNN implementation with multiple input images

Hello I am trying to implement a CRNN with multiple input images (in my context it is 6 images) This is a regression problem and output is two real value. And for the CNN block I am thinking of using ...
1
vote
2answers
58 views

How can I use 1-channel images as input to a CNN?

I need to develop a convolutional neural network whose inputs are 1-channel images, but I dont know how to do it, given that most libraries use 3 channel images. Should I convert my images to RGB? Is ...
1
vote
0answers
48 views

How High and Low frequency filters effect activation in the next layer?

Generally, we come across terms such as High Frequency and Low frequency filters in Convolutional Neural Networks (CNN). In regards to this highlighted statement, in 'S1' section of this paper by ...
1
vote
1answer
17 views

Convolutional Neural Networks for different-sized Source and Target

CNNs are often used in one of the following scenarios: A known-sized image is encoded to an intermediate format for later use An intermediate or precursor format is decoded into a known-sized image ...
0
votes
0answers
17 views

Not clear about CoordConv

I read the CoordConv paper and I am a bit confused about its implementation for a GAN/VAE. I understand how to add 2 more channels to an image and pass that to a conv net (and there are good online ...
0
votes
0answers
50 views

Dense bottleneck layer in Autoencoder

I would like to use the bottleneck layer of U-Net (last layer of the encoder) to calculate the similarity between two images. For that I have to somehow flatten the last layer of the encoder. In my ...
2
votes
2answers
101 views

What is the use of softmax function in a CNN?

What is the use of softmax function? Why was it used at the end of fully connected layer in convolution neural network?
2
votes
1answer
34 views

How does ARKit's Facial Tracking work?

iPhone X allows you to look at the TrueDepth camera and reports 52 facial blendshapes like how much your eye is opened, how much your jaw is opened, etc. If I want to do something similar with other ...
0
votes
1answer
45 views

Why does Convolutional layer unde usually has the same input/output channel size?

As famous model VGG16 shows(and other famous models), The convolutional layers before pooling usually have the same input and output channel sizes? What's the reason for that? Is there a theory or ...
0
votes
1answer
115 views

Autoencoder for MobileNetV2

I have way more unlabeled data than labeled data. Therefore I would like to train an Autoencoder using MobileNetV2 as the encoder. Then I will use the pretrained model for the classification of the ...
1
vote
1answer
73 views

Why should each filter have different weights for each input channel?

From the answers to this question In a CNN, does each new filter have different weights for each input channel, or are the same weights of each filter used across input channels?, I got the fact that ...
2
votes
0answers
34 views

Can GANs be used to generate matching pairs to inputs?

I have some limited experience with MLPs and CNNs. I am working on a project where I've used a CNN to classify "images" into two classes, 0 and 1. I say "images" as they are not actually images in the ...
0
votes
1answer
54 views

What is wrong with this CNN network, why are there hot pixels?

I'm building a CNN decoder, which mirrors (in reverse) the VGG network structure from Conv-4-1 layer. The net seems to be working fine, however, the output looks broken. Please note that the colour ...
3
votes
0answers
29 views

How do we give a kick start to the Facenet network?

I read the Facenet paper and one thing I am not sure about (it might be trivial and I missed it) is how do we give the kick start to the network. The embeddings in the beginning are random, so ...
0
votes
0answers
47 views

Convolutional neural network debugging

Im trying to implement CNN for small images classification (36x36x1) (grayscale). I've checked every forward/backward pass function on small example, and still my cnn is not doin any progress on ...
2
votes
1answer
35 views

Neural Nets: CNN confirming layer/filter arithmetic

I was hoping someone could just confirm some intuition about how convolutions work in convolutional neural networks. I have seen all of the tutorials on applying convolutional filters on an image, but ...
0
votes
0answers
55 views

Super Resolution on text documents

I want to implement super-resolution and deblurring on images from text documents. Which is the best approach? Are there any Git-hub links which will help me to start? I am new to the field. Any help ...
0
votes
1answer
105 views

How can I train a deep learning model to predict a matrix?

I am trying to train a deep learning model to predict an 8*2 matrix. The predicted matrix would have complex values and the input matrix would be real numbers. Can it be done? Thank you for your time.
3
votes
1answer
57 views

Can a CNN be trained incrementally?

Like our human brain, we can first learn (train) the handwriting 0 and 1. After the traing (and test) accuray is good enough, we only need to study (traing) the hardwriting 2, Instead of cleaning all ...
1
vote
0answers
34 views

Object size identification and maximum number of classes with convolutional neural networks

I am working on a project that involves using a ConvNet to identify screws. I am able to train from scratch a ConvNet based on the first version of the inception network, but shallower (only 3 ...
1
vote
0answers
22 views

Training a reinforcement learning model with multiple images

I am tentatively trying to train a deep reinforcement learning model the maze escaping task, and each time it takes one image as the input (e.g., a different "maze"). Suppose I have about $10K$ ...
1
vote
1answer
41 views

How to add some data input in a CNN?

There is this problem I have encountered, I was trying to classify the pixels from input image into classes, sort of like segmentation, using a encoder-decoder CNN. The “interested” pixels usually ...
2
votes
1answer
42 views

Convolutional layer to Fully Connected Layer implementation

Im implementing a neural network framework from scratch in C++ as a learning exercise. There is one concept I don't see explained anywhere clearly: How do you go from your last convolutional/pooling/...
3
votes
2answers
52 views

Are filter kernels fixed or learned?

No matter what I google or what paper I read, I can't find an answer to my question. In a deep convolutional neural network, let's say AlexNet (Krizhevsky, 2012), filter weights are learned by means ...
1
vote
0answers
21 views

Learning utility function for AIS data

I am trying to learn utility functions for ships through their AIS data. I have a lot of data available and plan on focusing on fishing boats. So far I've researched a lot of IRL algorithms but I'm ...
0
votes
2answers
806 views

RTX 2060 vs GTX 1060 6GB for deep learning

First of all I don't play games at all and I still quite new to deep learning.I was using Alex-net(transfer learning actually) in MATLAB to classify images in my current laptop(i5-3230,without any GPU)...
1
vote
2answers
47 views

What are the value of the pixels of the convolved image?

I'm studying convolutional neural networks from the following article https://ujjwalkarn.me/2016/08/11/intuitive-explanation-convnets/. If we take a grayscale image, the value of the pixel will be ...
4
votes
1answer
117 views

How are filters weights updated for a CNN?

I've been trying to learn backpropagation for CNNs. I read several articles like this one and this one. They all say that to compute the gradients for the filters, you just do a convolution with the ...
1
vote
1answer
35 views

Convolution layer neurons when extracting multiple feature maps

I've recently been reading up on CNNs and this part of the architecture is really confusing me. Assume, I have an input of size [32*32*3] and pass it to a convolution layer. Now, if my kernel size ...
0
votes
2answers
64 views

How to manage large amounts of image data for training?

Right now, I am trying to synthesize training images for a CNN and due to the nature of the application, there is a finite number of sample images to learn from. From other research, I expect to be ...
2
votes
2answers
44 views

Is the number of feature maps equal to the number of kernels in the LeNet 5 architecture?

In LeNet 5's first layer, the number of feature maps is equal to the number of kernels. However, the second convolutional layer has a depth different from the 3rd layer. Does the filter size dictate ...
0
votes
0answers
6 views

Super resolution of an object in a video using adjacent frames

Video frames super-resolution with deep learning? I've been searching for the whole day and could find no papers\projects tackling that problem. For example: suppose I have a series of N frames of ...
2
votes
0answers
11 views

How to debug and find neurons that most influenced a pixel in the output image?

I'm building CNN network of Image to Image. After training, I have some bad results in part of the Image. I would like to find the neurons that most influenced those pixels and do retraining only ...
0
votes
1answer
76 views

Artifacts After pruning Unet CNN

Im trying to make a dark image brighter using CNN-UNet arcitecture. When I train the network I get the following results: When I cut the features in half for pruning, and do full train again, I get ...
0
votes
0answers
37 views

How would you feed a neural network a variable sized array as an input?

From what I've seen, neural networks take a set of atomic inputs. I want an input to be a variable array, i.e. a group of people (with unique IDs). If I didn't care about their ID, I could simply ...
0
votes
0answers
15 views

Bigger receptive field

I have a network which has a input size of (28x28x1) and since I'm using (3x3 convolution) so the receptive field is (3x3). Before going further I will show the code snippet ...
1
vote
0answers
17 views

Learning Features from a Pre-trained Network

I am currently working on learning the features provided by a pre-trained network for image retrieval. Currently I take the features provided by the pre-trained network, use global max pooling to ...
0
votes
0answers
44 views

Can we use Autoencoders for unsupervised CNN feature learning?

I searched through the internet but couldn't find a reliable article that answers this question. Can we use Autoencoders for unsupervised CNN feature learning of unlabeled images like the below and ...
1
vote
1answer
18 views

Convolutional Sequence to Sequence Learning kernel parameters

I am reading the paper Convolutional Sequence to Sequence Learning by Facebook AI researchers and having trouble to understand how the dimensions of convolutional filters work here. Please take a look ...
0
votes
0answers
23 views

Net stops to learn when I increase number of classes

I'm kind of stuck, and instead of trying to randomly shoot the net with my ideas maybe I can consult it with you (one epoch takes 7h, so I cant't test my random ideas). Here's the crime scene: My ...
1
vote
2answers
54 views

Is it possible for a NN to reach the same results as CNNs?

Can a normal neural network work as good as a convolutional network? If yes, how much more time and neurons would it need compared to a CNN?
1
vote
1answer
125 views

What are the various methods for speeding up neural network for inference?

One way to speed up a neural network is to prune the network and reducing number of neurons in each layer. What are the other methods to speed up inference?
0
votes
0answers
19 views

Parameterized Coordinates in Region Proposal Networks (RPNs) for Faster R-CNN

In the original Faster R-CNN paper, the authors parameterized the box coordinates for regression under RPN. Below is the snippet of how they computed it:        &...
0
votes
1answer
56 views

How can I suppress a CNN’s translation invariant or translation equivariant?

I am trying to understand this post, but I get confused by the definitions and the differences. What's definition of equivariant? If I remove all the pooling layers from a CNN, will it make the ...
1
vote
0answers
23 views

Point Cloud Alignment using a Neural Network?

Having two point clouds, the second being a transformation of the first, how could I utilize a neural network in order to solve the pose (transformation in terms of x, y, z, rx, ry, rz) of the second ...
1
vote
0answers
27 views

Simple weakly supervised Object localizetion using keras. How to visualize the results?

I am following this link : Weakly-supervised-object-localization to create heatmap of the region in an image where the CNN looks to identify the class. As per the above mentioned repository , ...
1
vote
1answer
39 views

Does changing the order of the convolution layers in a CNN have any impact?

Could changing the order of convolution layers in a CNN improve accuracy or training time?
-1
votes
1answer
57 views

Use Image data from Drive to Colab for Image Augmentation

I am working on CNN. I have saved images in drive to do image augmentation in keras, I have used method (.flow_from_directory(directory)) Since it require directory path. I have mounted drive and give ...