Questions tagged [convolutional-neural-networks]

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

Filter by
Sorted by
Tagged with
1
vote
2answers
41 views

Do bounding boxes increase accuracy in and of themselves?

Say I have a standard image classification problem (ie: CNN is shown a single image and predicts a single classification for it). If I were to use bounding boxes to surround the target image (ie: ...
0
votes
2answers
73 views

How should the values of the filters of a CNN change?

I wrote a convolutional neural network for the MNIST dataset with Numpy from scratch. I am currently trying to understand every part and calculation. But one thing I noticed was the "just positive" ...
0
votes
1answer
41 views

How to identify the areas to reduce over fitting?

I am trying classify CIFAR10. The CNN that I generated over fits when the accuracy reaches ~77%. The code and the plot is given below. I tried DropOut, Batch Normalization and L2 Regularization. But ...
2
votes
1answer
30 views

how to benefit from previous training weights in training again to increase accuracy?

I have trained a modified VGG classification CNN, with random initialized weights; therefor the validation accuracy was not high enough for me to accept (around 66%). now using the weights resulted ...
3
votes
1answer
103 views

Variational Autoencoder task for better feature extraction

I have a CNN with the regression task of a single scalar. I was wondering if an additional task of reconstructing the image (used for learning visual concepts), seen in a DeepMind presentation with ...
2
votes
1answer
72 views

How do I choose the number of neurons in the fully-connected layer before the softmax layer?

I am solving a classification problem with CNN. The number of classes is 5. How can I decide the number of neurons in the FC layer before the softmax layer? Is it $N * 5$, where $N$ is the number of ...
3
votes
1answer
56 views

Confused about group convolution

I think I don't understand group convolutions well. Say you have 2 groups. This means that the number of parameters would be reduced in half. So assuming u have an image and 100 channels, with a ...
4
votes
1answer
70 views

Is it possible to make a 'forked path' neural network?

I want to make a network, specifically a CNN for image recognition, that takes an input, processes it the same way for several layers, and then at some point splits before coming to two different ...
1
vote
0answers
64 views

What are the key differences between cellular neural network and convolutional neural network?

What are the key differences between cellular neural networks and convolutional neural networks in terms of working principle, implementation, potential performance, and applicability?
3
votes
1answer
113 views

Why are not validation accuracy and loss as smooth as train accuracy and loss?

I am training a modified VGG16 network for classification (adding 0.5 dropout after each of the last FC layers). In the following plot I am training for a small number of epochs as an example, and it ...
3
votes
0answers
38 views

How do I denoize a microscopic image?

I'm working in a computer vision project, where the goal is to detect some specific parasites, but now that I have the images, I noticed that they have a watermark that specifies the microscope ...
0
votes
0answers
15 views

Using the cloud service to trasform a picture using a neural algorithm?

yesterday I tried to transform a picture in the artistic style using CNNs based on A Neural Algorithm of Artistic Style by Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge using a recent Torch ...
1
vote
1answer
49 views

When is bias values updated in back propagation?

I am new to deep learning. I have doubts on modifying bias values during back propagation. My doubts are Does the back propagation algorithm modifies the weigh values and bias values in the same pass?...
0
votes
0answers
44 views

CNN doesn't learning simple geometric patterns

It must be a very stupid question, but since I have not such sufficient know ledge storage and having no more time to search the answer of it, I have to put it here to ask for help. I generated a ...
0
votes
1answer
33 views

CNN output generally has more than one category in one-hot categorization?

I'm a bit of a CNN newbie, and I'm trying to train one to image classify pictures of pretty similar looking particles. I'm making the inputs and labels by hand from a set of 48x48 grayscale images, ...
2
votes
0answers
40 views

Algorithms to indentify people in pictures without using face recognition

There are lot of researches about face detection in pictures, but is it the only way one can say "this person I'm looking for is here in this picture"? Aren't there algorithms that you can provide ...
2
votes
1answer
28 views

Extending a neural network to classify new objects

Suppose a model M classifies apples and oranges. Can M be extended to classify a third class of objects, e.g., pears, such that ...
0
votes
1answer
141 views

Back propagation on Flatten Layer in CNN

I am making a NN library without any other external NN lib and is implementing the Flatten layer. I know the forward implementation of flatten layer but is the backward just reshaping it or not? If ...
0
votes
0answers
31 views

Which neural network to use for mapping a vector of size m to a vector of size n, where n >> m?

I am trying to solve a mapping problem on a grid (100x100) where I have few points, say 10, where I know the values of a tensor $\boldsymbol{M}$. I have a scalar field, $v$, which is related to the ...
0
votes
0answers
45 views

Double Convolution Layers in Yolov3

Lately, I have been working on yolov3 and have been trying to train it on x-ray images to detect a fracture. However, I have decided that I would want to increase the number of convolution layers for ...
3
votes
1answer
113 views

What is a simplified way to explain why the AI researchers Bengio, Hinton, and Lecun, won the 2019 Turing Award?

The Turing award is sometimes called Computer Sceince's Nobel Prize. This year's award goes to Bengio, Hinton, and LeCun for their work on artificial neural networks. The actual work contributed by ...
1
vote
0answers
23 views

Classification of classes within meta-classes

TLTR: I'm developing a CNN for a classification task. The data contains multiple classes some of which are very similar to each other and I know these meta-classes. In such a situation is it a good ...
2
votes
0answers
63 views

Convolutional Sequence to Sequence Learning: Training vs Generation

I am struggling to understand the use of the Convolutional Sequence to Sequence (Conv-Seq2Seq) model. The image below is take directly from the paper and is the nearly canonical diagram of the ...
1
vote
0answers
28 views

how to add the pool4 to the 2 x conv7 in FCN-16s using keras?

Now I'm using tensorflow.keras to implement the FCN-16s, this picture may be different with others, you should focus this it add ...
1
vote
0answers
14 views

One dimension deconvolutions or fully connected layers?

I’ve created a variational autoencoder to encode 1-dimensional arrays. The encoding is done through 3 1d-convolutional layers. Then, after the sampling trick, I reconstruct the series using 3 fully ...
1
vote
2answers
51 views

Detecting abnormalities in x-rays while taking into account demographics of a patient -automated

This is my first post so please forgive me for any mistakes. I am working on an object detection algorithm that can detect abnormalities in an x-ray. As a prototype, I will be using yolov3 (more ...
1
vote
0answers
31 views

Image-Specific Class Saliency Visualisation

In the paper "Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps", https://arxiv.org/abs/1312.6034, at part 3, there is a first-order Taylor expansion(...
5
votes
2answers
90 views

How do I improve accuracy and know when to stop training?

I am training a modified VGG-16 to classify crowd density (empty, low, moderate, high). 2 dropout layers were added at the end on the network each one after one of the last 2 FC layers. network ...
0
votes
0answers
29 views

Can we compare MAE MSE results with categorical_crossentropy?

can i compare MAE and MSE loss results of a regression CNN with categorical_crossentropy loss of a classification CNN if they both have similar tasks? is yes how to?
0
votes
0answers
35 views

How to Add Spp(Spatial Pyramid pool) layer to CNN network?

i create A model based on ELA [error level Analysis] for image forgery detection i use the following code : ...
1
vote
1answer
45 views

How to chose dense layer size?

I am fine-tuning a VGG16 model on 20 classes with 500k images I was wondering how do you chose the size of the dense layer (the one before the prediction layer which has a size 20). I would prefer not ...
1
vote
0answers
20 views

Very Large 3D Input Size

I would like to use a 3D convolutional network on a 2000x2000x2000 volume for segmentation. I know I can break the volume into chunks that can fit in VRAM, but I was wondering if there was a way to ...
0
votes
0answers
16 views

how to work with multi-labels or two inputs and a output

I’m in this problem and haven’t found a sound solution to it. Been like 20 days now. I have a dataset that looks like this: ...
0
votes
0answers
26 views

metrics evaluation multiclass classification

I am working on intent classification task (chatbot engine), 2k sentences, 24 classes. Major class is composed of about 150 sentences, minor class of about 35 sentences, the others are more or less ...
1
vote
0answers
26 views

Is making lot of 1 versus other model efficient?

I've got classification problem on image, I have 10 classes and when I fine tuned my model on it (I tried VGG, Xception, resnet etc) I have approximatly 83% validation accuracy. I was wondering if ...
1
vote
1answer
38 views

Why don't we perform classification of crowd density?

For the case of crowd density estimation using CNN, using datasets like shanhaiTech or UCF, why there hasn't been attempts to tackle this type of task as a classification problem? All current papers I'...
1
vote
0answers
14 views

any workaround to manipulate/transform recurrent CNN for sentence classification?

I learned how to build recurrent cnn model for text classification and sketched out my initial implementation. However, I am wondering how to transform recurrent <...
1
vote
1answer
57 views

Why doesn't my image classification network get better with training?

I am attempting to train a network to do something I thought would be a relatively simple case to learn with: identify whether the back of a scanned vintage postcard has one of 'no postage stamp', a '...
0
votes
0answers
23 views

Maximum number of nodes for CNN for mobile phones

I am building a mobile camera app that needs to show a real time preview of the processed image. So i was thinking of reducing the image size to 100 x 100. But as for the weights and number of layers. ...
0
votes
1answer
46 views

How to reduce over-fitting on training set?

Currently I'm feeding spectrogram of audio to the CNN with 3 convolution. Each convolution is followed by a max pool of filter size 2. First -> 5x5x4 Second - > 5x5x8 Third - > 5x5x16 and final ...
2
votes
0answers
39 views

How can the convolution operation be implemented as a matrix-vector multiplication?

How can the convolution operation used by CNNs be implemented as a matrix-vector multiplication? We often think of the convolution operation in CNNs as a kernel that slides across the input. However, ...
0
votes
0answers
31 views

How is the convolution operation used in CNNs a special case of the convolution operator?

How is the convolution operation used in convolutional neural networks (CNNs) a special case of the mathematical convolution operator? Most of us, when we think of the "convolution operation", we ...
1
vote
0answers
68 views

Siamese Network for unknown object

I am currently trying to create a One-Shot network using the Siamese architecture for an object that isn't a face. My problem is, in normal Face Recognition the detecting gadget (e.g. Smartphone) ...
1
vote
1answer
339 views

Count number of objects in image using CNN

I'm looking for neural network architecture that excel in counting objects. For example, CNN that can output the number of balls (or any other object) in a given image. I already found articles about ...
0
votes
0answers
19 views

How does the process of segmentation of face in face recognition work?

How does the process of segmentation of face, using a CNN, in face recognition, work? How are we able to segment the face?
1
vote
2answers
402 views

Why does a fully connected layer only accept a fixed input size?

I'm studying how SPP (Spatial, Pyramid, Pooling) works. SPP was invented to tackle the fix input image size in CNN. According to the original paper https://arxiv.org/pdf/1406.4729.pdf, the authors say:...
0
votes
0answers
161 views

1D GAN not converging

I am trying to build a 1D GAN able to produce data similar to the input one, which looks like this: I am using pytorch. This is the code for my Discriminator, which takes as input a 1D vector of size ...
0
votes
1answer
47 views

How to study the correlation between GAN's input vector and output image

A generative adversarial network (GAN) takes a vector of numbers as input and generates an image, based on the input. Each element of the vector causes some feature of the image to change, but the ...
1
vote
1answer
16 views

Using convnet to classify language of text contained in images

I hope this question is not too broad or general. I have a very large set of images all of which contain text (some have more, some less). All of them have been tagged as containing, say, English text ...
2
votes
0answers
24 views

Calculating tangent vector of curve s(P,$\alpha$) at given point $\alpha$ = 0

I am reading the paper "Transformation Invariance in Pattern Recognition – Tangent Distance and Tangent Propagation", where the tangent vector is calculated for the given curve $s(P,\alpha)$ at $\...