Questions tagged [conv-neural-network]

Convolutional Neural Networks are a type of neural network in which only subsets of possible connections between layers exist to create overlapping regions. They are commonly used for visual tasks.

Filter by
Sorted by
Tagged with
0 votes
0 answers
7 views

How can I train Mixture Density Neural Network?

I am learning Mixture Density Neural Network but it looks different from the usual neural network for regression problem. As far as I have understood from what I have read on the Internet, it gives ...
user avatar
  • 2,585
1 vote
0 answers
23 views

How to deal with unknown classes with a convolution neural network classifier?

I'm quite new into the DL and ML field. I'm training a CNN able to classify 3 different classes, however I would like in the testing phase to make the CNN able to not misclassify images that do not ...
user avatar
  • 11
1 vote
1 answer
18 views

Concatenation or separate channels for a CNN

let's say I am classifying time series data from multiple channels in a biomedical setup (e.g. 12 lead ECG). I have been reading this paper on a CNN-based (ResNet) architecture for assesing the ...
user avatar
  • 111
0 votes
0 answers
17 views

Best way to approximate head point having only face keypoints

I'm using the BlazeFace model from TensorFlow which only has this few keypoints: I need those keypoints plus a head keypoint, like this one: My question is, which would be the best way to ...
user avatar
0 votes
0 answers
63 views

My CNN is not learning but just memorizing [duplicate]

So there has been similar posts but none of them solves my problem, so I decided to created a new question. I'm working on a regression project where I intend to use CNN to predict material properties,...
user avatar
  • 1
0 votes
0 answers
7 views

implementing a neural network U-Net with imbalanced classes, implementing the loss function

my problem is : i have a neural network U-Net, but to do the segmentation on my sparse annotation, i need to implement the loss function for the imbalanced classes so the article says, that there is a ...
user avatar
0 votes
0 answers
6 views

How to calculate number of learnable parameters in CNN? [duplicate]

How to calculate number of learnable parameters in CNN when only kernel size and number of filters are given? Lets say the kernel size is x and number of filters is y. In that case in which way I ...
user avatar
1 vote
1 answer
36 views

Does having the same background for all the images of a particular class increase a CNN model's ability to classify the images?

I'm working on a multi-class skin disease classification problem. The input images are skin diseased images that have varying backgrounds. Does maintaining the same background (maybe a white/black ...
user avatar
  • 53
4 votes
1 answer
39 views

Does watermark/text on images at the same position influence the classification of images using CNN?

I'm working on a multi-class classification problem using CNN. Most of the images of each class have a text/watermark at a specific position on the image. I have a couple of questions. Does the ...
user avatar
  • 53
0 votes
0 answers
13 views

What are some of the commonly used image processing techniques for multiclass image classification?

I'm working on multiclass skin disease image classification(caused by bacteria and fungus). Some of the sample images are shown below. Images contain different background as shown in image_1 and ...
user avatar
  • 53
0 votes
0 answers
9 views

Strategy for Train/Test-Split on Video Sequences

My dataset consists of 15 video sequences, each sequence showing a different movement. I want to train a CNN to detect poses (e.g. standing, sitting, ...) on single frames of this dataset but struggle ...
user avatar
1 vote
0 answers
38 views

Regular *negative* spikes in neural network training loss

I am training an ALL-CNN network using the Adam solver. As the figure shows, the testing seems to converge to an acceptable solution, but there are these regular negative spikes during training that ...
user avatar
1 vote
0 answers
52 views

Derive the gradient matrices w.r.t. W1 and W2 and backprop equation in a Residual Network [closed]

How would I go about deriving gradient matrices w.r.t. W1 and W2 and backpropagation equation in a residual block that is a part of a larger ResNet network with forward propagation expressed as: $$ F(...
user avatar
  • 11
0 votes
0 answers
10 views

Why does MobileNet Architecture start with a Standard Convolution?

I am trying to understand the design choices behind the MobileNet architecture. (pdf available on the right). The authors use Depthwise Separable Convolutions as a replacement for Classical ...
user avatar
  • 1
1 vote
1 answer
14 views

Can self-supervised pretraining work with only labeled data?

I am working on an image classification problem with only a few samples (10 images). As part of the challenge, we aren't allowed to use any external data or pretrained models. I was wondering whether ...
user avatar
0 votes
0 answers
28 views

How am I supposed to choose model, based on the best validation set accuracy or last epoch?

I trained the VGG Network on cifar10. Use my customized noise label dataset and clean dataset, then divide them into train, validation and test sets. In clean dataset everything went well, high ...
user avatar
1 vote
1 answer
23 views

In a multioutput deep learning model, is there a benefit to normalizing the output dimensions if they are of different magnitudes?

I am building a multi output deep learning model where the output consists of five dimensions (the specific architecture is a modification of YOLO). These have different magnitudes (ranges: [0, 1.2], [...
user avatar
1 vote
1 answer
65 views

Implementation of a convolution layer in a cnn

This questions deals with the implementation of a convolution layer. First I like to make clear what I understand about cnn's. A cnn uses filters/kernels to find geometrical features from the input ...
user avatar
  • 11
1 vote
0 answers
11 views

"Making VGG-style convNets Great Again" - Reparam section unclear [closed]

I have been reading the article: RepVGG: Making VGG-style Convnet Great Again I have reached the section 3.3 Reparam for plain Inference model and it's unclear for me.I have read this section dozen ...
user avatar
0 votes
0 answers
6 views

Interpreting validation loss and accuracy for various learning rates

I am having a hard time comparing the effect of different learning rates on validation loss and accuracy. Would I be right in assuming that a Learning rate of 0.0001 was the most successful as the ...
user avatar
0 votes
0 answers
19 views

If the weight and bias gradients are stuck at zero throughout training, is this an indication of dying ReLu?

A high learning rate when combined with a ReLu activation function is known to lead to the 'dying ReLu' problem. Is this a reasonable conclusion to arrive at if the gradient with respect to weights ...
user avatar
2 votes
1 answer
34 views

How to count the parameters in a convolution layer?

I'm preparing for an exam in Computer Vision. I came across with the following question from one of the exams: What is the number of parameters of a convolution layer in a neural network, when the ...
user avatar
  • 161
0 votes
0 answers
3 views

Can I use low dimensional node features in graph convolutional networks?

I am trying to understand how GCNs work. For example, the well known GraphSAGE algorithm considers a graph $G$ with node features $x_i$ of dimension $n$. Then it propagates the node features over the ...
user avatar
0 votes
0 answers
14 views

FID as a metric to evaluate the quality of synthetic datasets (Non GAN generated) for training models for a given classification task

I am working on a problem of generating synthetic data (algorithmically by blender, not using GANs) to aid the training of some CNN for a classification ask. Ideally, I want to generate an algorithm ...
user avatar
  • 135
0 votes
1 answer
23 views

Is there such a thing as intra-sample modal collapse in GANs?

Mode collapse is a known issue in generative adversarial networks (GANs) whereby the generator only learns a subset of the real data distribution. In those cases, it only outputs variations of a small ...
user avatar
1 vote
1 answer
32 views

Can anyone tell me, how should I test out model? should I test it with each epoch or should we test at the end?

Lets Say, here is our train and test function where we train and test the model. ...
user avatar
0 votes
0 answers
21 views

How can I mathematically explain the convolutional neural network?

How can I give notations to the whole CNN network? Like, if I feed input 'x' to feature extractor, we will get the extracted feature embedding vector at the end that will be passed to the linear ...
user avatar
0 votes
0 answers
13 views

Are there any statistical principles that a neural network layer should obey?

In the neural networks (NNs), we have different layers. Probably the best resource on finding zoo of different layers is the Keras Layer API. Some of them used as only passing and reshaping. Most of ...
user avatar
  • 1,816
1 vote
1 answer
52 views

How do CNN filters learn from back-propagation?

I have some intermediate knowledge of Image-Classification using convolutional neural networks. I'm pretty aware to concepts like 'gradient descent, 'derivatives', 'back-propagation & 'weight ...
user avatar
0 votes
0 answers
16 views

CNN model accuracy is increasing really slowly [duplicate]

I am training a CNN model from scratch on the Caltech101 dataset. The accuracy of the model is increasing very slowly after the 5th epoch. Shown below is the accuracy and loss curves of the model for ...
user avatar
1 vote
0 answers
30 views

Why is the convolutional filter flipped in convolutional neural networks?

I don't understand why there is the need to flip filters when using convolutional neural networks. According to the lasagne documentation, flip_filters : bool (default: True) Whether to flip the ...
user avatar
2 votes
0 answers
78 views

Is CNN shift invariant?

I have seen multiple resources on the web says that CNN is or is not shift invariant. I do think that CNN is shift invariant as the convolution kernel will scan the image to detect the features that ...
user avatar
  • 231
0 votes
2 answers
31 views

Weight's shape of Convolution Neural Network?

I've have read that cnn have neuron per pixel but also read that it is not true. so what is the actual answer? and what I know is cnn tries to adjust the weight matrix which is also a kernel matrix, i ...
user avatar
0 votes
0 answers
8 views

Content-based image retrieval for a set dataset of tv shows

I'm trying to build something that will, given an image, find which tv show, episode, and timestamp it's from. I believe this is considered content-based image retrieval. It's sort of like Shazam (...
user avatar
1 vote
0 answers
49 views

DNN architecture for regression with small inputs, large output

TL;DR what DNN architecture for 1xn (n<50) tensor to 256x256x6 tensor regression? I'm facing a supervised regression problem where I need to predict the outcome of the numerical simulation of a ...
user avatar
3 votes
1 answer
36 views

Convolutional neural network architecture calculation question

I'm attempting to understand the neural network architecture used in this paper: Visualizing and Understanding Convolutional Networks. Here's an image of the network achitecture from the paper that is ...
user avatar
0 votes
0 answers
8 views

How do you scale the activation function of an auto-encoder when using a custom normalization fitted on the data?

I'm working on a convolutional auto encoder. The input is an image The output is a reconstructed image During the training phase, we feed the same image in and out The loss is the Mean Squared Error ...
user avatar
  • 111
0 votes
0 answers
15 views

What prevents my PyTorch convolutional auto-encoder to converge on some initializations? [duplicate]

I built a small auto-encoder for greyscale images. It is there to make some tests, so I train it often, and I have a strange behavior. On some initialisations, it does not converge. I mean, the MSE ...
user avatar
  • 111
2 votes
0 answers
52 views

Good performance on both training set and validation set, but poor performance on the test set

I'm building a CNN to classify the American Sing Language fingerprints. During training and validation it works very well, with accuracy going up to 95% and loss down to 10% on both. However, when it ...
user avatar
2 votes
2 answers
127 views

CNN - upsampling backprop gradients across average-pooling layer

How to up-sample gradients, during back-propagation, across an average-pooling layer? For this purpose, let $$ A^{[l]} = \begin{bmatrix} a_{11} & a_{12} & a_{13} \\ a_{21} & a_{22} & ...
user avatar
  • 206
0 votes
0 answers
44 views

Why are Convex Loss Functions important in SGD? [duplicate]

Why are convex loss functions so important in Neural Networks? Because Neural networks are learnt end-to-end, with non-linear activations, causing convex loss functions to actually become non-convex ...
user avatar
  • 252
2 votes
0 answers
19 views

How is it ensured that a CNN does not learn the same filter multiple times?

Let's say we have a dataset of pictures displaying straight vertical lines with different translations or nothing (just an example), and we apply a CNN to that. We choose that the CNN should learn N ...
user avatar
0 votes
0 answers
12 views

How is derivative calculated for Grad-CAM if the final output is multidimensional?

For Grad-CAM, the derivative of the final output is found with respect to the elements of the channel considered Selvaraju et al. 2019. But if the output is a multidimensional matrix how is the ...
user avatar
0 votes
1 answer
28 views

Complexity of 1D CNN and 2D CNN

Are the computational complexity of 1D CNN and 2D CNN the same? If not what are their computational complexity and what is the best way to compute them? Considering both forward and backward ...
user avatar
  • 101
1 vote
0 answers
14 views

How are the region proposals processed in Fast R-CNN during training and testing?

How are the region proposals processed in Fast R-CNN during training and testing? From my understanding, during training we sample some small number (e.g. 64) of region proposals out of the ~2000 ...
user avatar
  • 31
1 vote
0 answers
56 views

Is it okay to not use batchnorm and relu before global average pooling?

I have built and experiment with a small network by batchnorm-relu-conv rather than conv-batchnorm-relu as suggested by DenseNet(2017). In denseNet, Before global average pooling layer, there are ...
user avatar
  • 11
2 votes
0 answers
16 views

Convolution Backpropagation

My question is related to this question answered here. I want to find $\frac {\partial E}{\partial K}$. For this I do convolution between $\frac {\partial E}{\partial O}$ and $X$. I couldn't find ...
user avatar
1 vote
0 answers
75 views

How to apply Keras Conv1D over 3D dimensional input?

Context: I'm predicting whether a machine will break down within 1 hour, and I have sensors located at 4 different parts of the machine, which give me historical readings of different metrics. ...
user avatar
0 votes
0 answers
21 views

Can Grad-CAM be used for CNNs with a flatten layer?

I would like to visualize CNNs with a flatten layer. I looked into Grad-CAM, which is one of the most popular visualization methods for CNNs, but I thought it could only be used for CNNs with a global ...
user avatar
  • 1
1 vote
1 answer
62 views

"Milder" than a convolutional neural network: not forcing connections to be perfectly equal or exactly zero, but penalizing such behavior

Convolutional neural networks (CNNs) do regularization (of sorts) by forcing some weights to be dropped and others to be zero. Borrowing some drawings from another post of mine... Apply the filter to ...
user avatar
  • 31k

1
2 3 4 5
29