Questions tagged [cnn]

Convolutional Neural Networks (CNN, also called ConvNets) are a tool used for classification tasks and image recognition. The name giving first step is the extraction of features from the input data.

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

How to make a pipeline for Videos Dataset for TensorFlow [Sequence Matters] & train Model Effectively with Low Memory System

I am working on a Deep Learning project and I am facing an issue with the size of the dataset. I want to make a pipeline for video dataset [Sequence Matters]. Because if I try the load the whole ...
0 votes
0 answers
18 views

CNN resize images

Reducing images size will cause a loss of information for sure. If a have a model that perform better on resized images (50x50) than on original size images (224x224), what can I deduce ? There is a ...
0 votes
0 answers
7 views

The val_accuracy is higher than training accuracy, and the test accuracy is very low compared to both val_accuracy and train_accuracy

I am training a CNN model where, Training data=687 , validation data=102 , testing data=79 The validation accuracy is higher than training accuracy The test accuracy is very low compared to both ...
0 votes
0 answers
14 views

Strange results using UNet for image restoration

I am trying to use U-net type network structure on image restoration type task. However my results have strange color with check board artifact (As shown in the third image). Any advice? This is my ...
-1 votes
0 answers
12 views

Counting the number of parameters in CNN

I could not find the answers of the following questions. Q.1. You have an input volume of 32 x 32 x 3. You want to process time-series data with a 1D CONV with zero padding, stride of 1, and 2 filters ...
0 votes
0 answers
12 views

Performance metric of convolutional filter

Suppose that there are two identical CNN ( CNN1 and CNN2). If I changed a set of filters in CNN1. Then, the filters in CNN1 and CNN2 will make different features with identical input. Here, I want to ...
  • 1
1 vote
1 answer
29 views

Improve CNN classification accuracy

I am training a CNN model with about 20.000 images with two classes each 10.000 images. The size of the images vary between 50*50 pixel and 1000x500 pixels. I am resizing all images to the average ...
  • 113
1 vote
0 answers
14 views

What data can I obtain from this H5 file?

I created a CNN model and it is saved in h5 format. I used the Netron app, where I obtained the model architecture, however batchnormalization was not present. Is ...
0 votes
0 answers
30 views

CNN age classification --- low accuracy

I have a dataset of 34k (200x200) images and I want to build an 8 class age detector. I've tried a lot of different networks design, regularizations, dropout layers, grayscale images, data ...
0 votes
1 answer
17 views

Reusing a model, pretrained on 19 classes, for just one of those classes

I have a pretrained net for semantic segmentation, which has been trained on the cityscapes dataset and its 19 classes (Person, car, traffic sign, …). One of those is "Person". I am only ...
  • 1
0 votes
0 answers
12 views

Degree of freedom for NLP DL models

How degree of freedom can be estimated for NLP use cases where put is high dimensional vector (let us say word2vec used and dim size is 500) say and RNN or 1D CNN is used for modeling?
0 votes
0 answers
15 views

how to reduce the loss and improve the gradient flow - CNN

I am trying to improve this situation, in image classification[3 classes, softmax in the last layer], I constructed the neural network having 7[conv2d+Batchnormalization] layers + 1 linear layer, ...
0 votes
0 answers
9 views

Autoencoder on spectral image ASCII data file

I am discovering Autoencoder CNN, I took a look around on tutorials, I see lot's of examples with images as input. I was wondering if it is possible to use files whose pixels are represented by Ascii ...
0 votes
0 answers
9 views

Best layout for performing CNN with multiple images

I'm looking to create a CNN (probably using ResNet) for automating image rating of videos by processing a predetermined number of frames from the video (60 for example evenly distributed through the ...
0 votes
0 answers
29 views

Different results between training and evaluation phase on the same data

I have trained a CNN and in the training phase I obtained an accuracy of 36.5%. If I call model.predict() on the same test data of the training phase I only obtain ...
0 votes
0 answers
10 views

A dataset specificaly crafted to fail CNN

Is there a dataset of images that is specifically crafted in a way that it is reasonably easy for a human, but notoriously difficult for a CNN to learn? A couple of examples are The set of dogs on the ...
  • 111
1 vote
0 answers
29 views

Regression for arrays rather than numerical values

I have a dataset with a small amount of points (~500) and 10 features. The points are related to eachother, so that information about point 1 should help predict information about point 2 etc etc. ...
  • 11
0 votes
0 answers
14 views

Increase and decrease in depth of CNN blocks

Please see the image below as an example. To my understanding, the 1nn image is converted to 500*n~ * n~ with the help of 500 kernels applied on the same image. What I am confused about is how in the ...
  • 21
0 votes
0 answers
19 views

Format 2d numpy arrays for 2D CNN

I have a dataframe where each row contains the 2d numpy array of an image (200x200) like: Given this data format, how do I convert this data into a form where I can use a TensorFlow CNN model on it? ...
0 votes
1 answer
10 views

Difficulty loading data/running model on custom dataset derrived from DNA sequence data - TypeError when attempting to run model

I am a student who has some limited experience with keras, and for a new project recently decided to learn how to use pytorch to implement my models. I'm a beginner with both, so apologies in advance ...
0 votes
0 answers
12 views

how to keep correspondence between ROI and background of a particular image during feeding them as input to a network

I am working on a CNN. I would like to provide the cropped object (object of interest in an image) and its corresponding background as the input to the CNN. So, how can I provide a specific background ...
0 votes
0 answers
28 views

My loss remains at ~0.4 regardless of any changes

I am crossplotting my problem as suggested by a member of SO. I am quite new to CNN and ML in general, and I remain in a phase where I use it as a tool (instead of building my own network). I am using ...
0 votes
0 answers
23 views

How ca I reshape y_train , y_validation from train_generator?

I retrained ResNet-50 for iris flower classification in tensorflow using the following code: ...
  • 115
0 votes
0 answers
43 views

Classification report for binary class problem

I am building a binary CNN using single neuron in the last dense layer and "binary_crossentropy" as the loss function while compiling to predict either class1 or class2. I am having problem ...
  • 11
0 votes
0 answers
15 views

Unexplainable (for me) peak Batch-Entropy loss in FNN training

I've just read an article "Improving the Trainability of Deep Neural Networks through Layerwise Batch-Entropy Regularization" which introduces the notion of layer batch-entropy, states that ...
  • 101
0 votes
0 answers
11 views

How does a CNN work in detecting absence of features/objects?

I'm trying to understand how a CNN operates internally. Let's say I'm doing binary classification with 1 output neuron and a sigmoid to classify dog vs no dog. No dog meaning the image does not ...
  • 33
0 votes
1 answer
27 views

How to use strong labels in image classification?

I have a dataset where I have localized pixel-level annotations of a dataset of cancer vs non-cancer. Which deep learning methods can I use to optimize the model to focus on the localized regions of ...
  • 33
0 votes
0 answers
6 views

Point to plane distance as 3DCNN loss?

I want to train a model that regresses a plane/slice from a 3D volume (vol). I am defining the target slices/planes with 3 vectors: center position vector (pos) and two unitary orientation vectors (o1 ...
1 vote
1 answer
17 views

How is ResNet different from FPN?

I'm learning more about different variations of deep CNNs. Based from my understanding, ResNet makes use of skip connections that's also somehow shaped like a pyramid or triangle? How is this ...
  • 11
0 votes
1 answer
56 views

Colab Pro+ CNN GPU memory saturation

I have a problem when executing jupyter notebook for CNN in colab pro+, to train a model with a size of 560664x48x48x1. normally the data is composed of images with a size of 48x48. I used 10 fold ...
0 votes
0 answers
12 views

checking correctness of predicted values of a CNN

I have trained a normal CNN to recognize patients with a disease or not. I print the predicted values for the test set, and get the probability of the various images of belonging to a class rather ...
0 votes
0 answers
12 views

Is 500 epochs too much for a CNN project?

I am working on a project where I need to train a model with a data set of 250 images. My epochs count is 500. Is that too much? Will it overfit? I did this because ...
0 votes
0 answers
11 views

Fine-tune network with fewer labels to increase accuracy

I'm trying to detect just dogs in input images. Would a pre-trained network on COCO significantly perform better if it was fine-tuned using COCO again (or another dataset) where all non-dog instances ...
  • 293
0 votes
0 answers
19 views

Pytorch Conv1d expected scalar type Long but found Float

I am attempting to train a CNN that applies a 1D convolution to a dataset of length 360 with entries being 2D arrays of size (3822, 4). I am encountering the following error: ...
  • 1
0 votes
0 answers
32 views

PyTorch Conv1D on 2D Input

I am designing a neural network architecture to apply 1D convolution to a dataset of size 360, and each input is represented as a 2D numpy array of shape (3822, 4). Here is the model: ...
  • 1
0 votes
0 answers
15 views

How to feed temporal image data to a 3D CNN?

I am using TensorFlow and Keras to build a 3D CNN, where the 3rd dimension is time. I have a dataset which contains evolving images with time (51 frames were taken). For testing, I took 10 simulations....
  • 1
0 votes
0 answers
15 views

Validation Accuracy plateaued and not increasing using CNN

I am using cifar10 dataset and below is the code that I am using. I think that the model is regularized but after around 0.70 of validation accuracy, it plateaus. Following are the graphs of loss and ...
  • 1
0 votes
0 answers
7 views

Application of eigenvalues, eigenvector, transposed matrix

Can you give me please some application of eigenvectors, eigenvalues and matrix transposition in data science? I guess for eigen-values/vectors it would be linear regression PCA and NLP, alongside ...
0 votes
0 answers
25 views

Masking pixels for a CNN

I'm trying to implement a CNN with RGB and depth images. But my depth images are a little sparse. So I would like to mask out those neighborhoods in the RGB image where the depth neighbors are empty. ...
0 votes
0 answers
26 views

The CNN binary classification model bad prediction after adding training data

I've trained my binary classification CNN on a limited, augmented data set, initially with good results. Later, I used the model on external data and added false positives to the training set. After ...
0 votes
0 answers
7 views

I do not understand normalization and standardization

in my ML Course at Uni, a big topic was normalization and standardization, but still - I don't really get why we do that. More specifically, I'm working on a (fairly complex) CNN to predict missing ...
0 votes
0 answers
13 views

Image recognition of specific animal drawings with body parts (tiny dataset)

I have an assignment for a class where I need to visualy detect specific animal drawing models, and extract the colors from such model after identifying the different body parts. My problem is that my ...
0 votes
0 answers
15 views

How to see Latency at layer granularity in a CNN

I am finding documents or an example that measure Latency at layer granularity in the AlexNet model. Please could share or tutorial for me.
1 vote
2 answers
36 views

Loss decreases, but Validation Loss just fluctuates

I've been trying to implement object detection using a CNN architecture like this: ...
1 vote
1 answer
76 views

High level-Low Level features in U-NET

Why do the first layers of U-Net or CNN generate low-level features? Why not the last layers? What is the logic behind getting low-level features at the beginning of architecture? And yes, high-level ...
  • 87
0 votes
0 answers
41 views

CNN model with images - 100% accuracy on validation and test sets with limited data?

I have a binary classification problem (e.g. if there is a human in the room or not) with a small dataset of images from a thermal camera. Originally, those were 7 videos, which I have converted into ...
  • 101
1 vote
1 answer
719 views

ValueError: cannot reshape array of size 36276416 into shape (96,227,227,1)

I am running my LeNet code with LFW, but when I run it, I am getting the following error message: Here is the code that it is getting the error ...
  • 11
0 votes
0 answers
8 views

CNNs and classifying images in different scales

Let's suppose we want to classify images of the temperature distribution on a metal sheet using CNNs, but the dataset available has images in many different linear scales. I'm using temperature as an ...
0 votes
0 answers
7 views

What is meant by softly selecting a matrices from a set of matrices?

How is the softmax layer applied to the 1x1 conv layer in order to soflty select the matrix from a set of matrix denoted by A. From my understanding of convolutions, implementing a 1x1 conv layer on a ...
0 votes
0 answers
15 views

Fakebert implementaion

I am trying to implement this architecture of fake bert for fake news detection, but I don't know how to feed the word embedding from Bert. Help, please.

1
2 3 4 5
27