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.
1,365
questions
0
votes
1
answer
24
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
2
answers
75
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
1
answer
732
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
1
vote
2
answers
391
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
511
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 ...
- 91
0
votes
2
answers
146
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
3k
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
1
vote
0
answers
17
views
Loss function for classification problem
So I'm working on a classification problem, I used convolutional neural networks to classify grayscale ECG beat images of dimension 200x200 (I had around 4000 images for each class in training and I ...
- 23
0
votes
1
answer
63
views
Why does the computation time drastically increase when increasing kernel size in a CNN?
I'm doing some experimentation with a CNN and have 2 conv layers with 32 and 64 filters respectively. I started out with 3x3 kernel sizes and noticed that when I increased it to 5x5, 7x7 etc the time ...
- 101
0
votes
1
answer
19
views
How to deal with different amounts of data every day?
I am doing a time series prediction task. There are different amounts of news headlines every day, and the goal is a binary prediction task to predict next day's stock movement.
The amount of ...
- 165
1
vote
0
answers
290
views
Averaging multiple train-test splits to estimate the performance with higher variability?
I have a small size data set and I want to assess the effect of a certain type of cases on the overall model performance. For example, is the model biased against people of a certain age group?
Using ...
1
vote
0
answers
223
views
Val Loss and manually calculated loss produce different values
I have a CNN classification model that uses loss: binary cross entropy:
...
- 201
0
votes
1
answer
41
views
Keras model with 3 input images giving wrong output
I have created a keras model that takes 3 images as input, passes them to individual CNN backbone(mobilenet_v2) and fuse the results from 3 individual streams. These fused outputs further goes through ...
- 1
1
vote
0
answers
101
views
Neural Network is non deterministic on validation
We have a regression problem we are trying to solve.
We are using Transfer learning by using resnet50 and adding a linear activation layer at the end of it.
each image input is 3 layers of synthetic ...
- 201
1
vote
2
answers
135
views
My CNN image classification model gives good predictions in all but 2 classes. What should I do?
I built a CNN image classifier for a dataset that contains 6 classes. The dataset is balanced in all 6 classes. After training, the model gives pretty good prediction accuracy in all but 2 classes. To ...
- 11
0
votes
2
answers
132
views
how does convolutional layer work? [duplicate]
I have one question regarding CNNs. If we take a single convolutional layer it can have multiple filters right? Are these filters all the same? Is a single layer made only to detect one feature? I am ...
- 23
1
vote
1
answer
37
views
Need term or method name for evaluation of CNN without ground truth using e.g. a regression model
I have the following problem,
I have trained a CNN and I can evaluate the network in-sample. I want to use the trained model for the class prediction of images for which I have no ground truth. ...
- 11
0
votes
1
answer
4k
views
AttributeError: 'Functional' object has no attribute 'predict_classes''
I am trying to use run a GoogLeNet code using FERET datasets. When I run the code, I get the following error message:
...
- 11
0
votes
3
answers
310
views
How to verify if the behavior of CNN model is correct?
I am exploring using CNNs for multi-class classification. My model details are:
and the training/testing accuracy/loss:
As you can see from the image, the accuracy jumped from 0.08 to 0.39 to 0.77 ...
- 63
0
votes
0
answers
812
views
Keras/Tensorflow Error: Specified a list with shape [3,1] from a tensor with shape [32,1]
I have been experimenting with keras/tensorflow to build up my confidence and am currently trying to build a LSTM model that forecast the price of a stock based on the price of the stock in the ...
- 1
1
vote
0
answers
37
views
Accuracy drops when adding a fully connected layer for dimensionality reduction to a ResNet50
I'm training a ResNet50 for image classification and I'm interested in decreasing the dimensionality of the embedded layer, in order to apply some clustering techniques.
The suggested dimension is ...
- 61
0
votes
1
answer
118
views
Using CNNs to detect incorrect label images in dataset
What I want to do is to train a model to identify the images that are incorrectly labeled in my dataset, for example, in a class of dogs, I can find cats images and I want a model that detects all ...
- 31
0
votes
1
answer
78
views
Data Augmentation Keras length of data
I'm confused when I add data augmentation should I get more data or the same data I tested my x_train length to confirm but I got the same length before augmentation and after augmentation is that ...
- 1
1
vote
1
answer
258
views
Overfitting CNN model - any relation to input image size?
If my CNN model is over-fitting despite trying all possible hyper parameter tuning, does it mean I must decrease/increase my input image size in the Imagadatagenarator?
- 53
0
votes
1
answer
50
views
Is it feasible to integrate convolutionnal layers as Reinforcement Learning input to learn video game?
Let's say, you want to apply reinforcement learning on a simple 2D game. (ex : super mario)
The easy way is of course to retrieve an abstraction of the environnment, per example using gym and/or an ...
- 3
1
vote
0
answers
1k
views
YOLOv5 can't detect object on custom dataset
Context: I'm trying to utilize an object detection model (YOLOv5) to detect damage/defects on cars (dents, scratches, cracks). Right now the goal is to make a minimum viable prototype, a model able to ...
- 45
0
votes
1
answer
131
views
What does it mean if the validation accuracy is equal to the testing accuracy?
I am training a CNN model for my specific problem. I have divided the dataset into 70% training set, 20% validation set, and 10% test set. The validation accuracy achieved was 95% and the test ...
- 145
0
votes
1
answer
502
views
How to compute the mean of weights of multiple models?
Hi i'm a student and i'm working on a Federated Learning problem, but before doing that with the proper tools like OpenFL or Flower, I started a little experiment to try in local to train using this ...
1
vote
1
answer
2k
views
My validation loss is too much higher than the training loss is that overfitting?
I am new to data deep learning. I am educating myself but I don't understand this situation. Where Validation loss is much much higher than the training loss. Can someone please interpret this?
...
- 13
1
vote
0
answers
68
views
Post processing in medical segmentation with attemtion unet
I am doing a lesion segmentation for multiple sclerosis (MS), and at the moment I am using a attention unet for my thesis. The best validation dice score I have recieved is 0.771 and train 0.84. I am ...
- 11
0
votes
0
answers
2k
views
Value Error: Shapes (None,128,128) and (None, 4) are incompatible
I am trying to perform CNN on my dataset. I came across the below error
ValueError: Shapes (None, 128, 128) and (None, 4) are incompatible
The shape of my xTrain ...
- 1
1
vote
1
answer
36
views
Creating a map between N images and N labels using CNN
I have seen classification CNNs that train with numerous images for a subset of labels (i.e. Number of images >> Number of labels), however, is it still possible to use CNNs when the number of ...
- 11
0
votes
1
answer
279
views
How to measure the similarity between two medical images of different imaging modalities according to similar objects in both of them?
I have two series of medical images each one from different imaging modalities.
According to that, I have been segmented the Region of interest (the object which appears in both modalities )using U-...
- 1
1
vote
0
answers
7
views
Hardware datapaths for weights and operands
A paper, Survey and Benchmarking of Machine Learning Accelerators, mentions
Conversely,
pooling, dropout, softmax, and recurrent/skip connection
layers are not computationally intensive since these ...
- 143
0
votes
1
answer
49
views
Am I over-complicating stuff?
I'm trying to classify some 1-D time series data, so I used a simple 1D CNN and fine-tuned the model via Bayesian Optimization (nothing fancy, just used the Keras tuner). And I got very good results (...
1
vote
0
answers
24
views
Trainning CNN Metric Learning gradient is constantly 1
I'm training a CNN with shirts and bodycon photos. I have these two classes and about 15k photos. I'am trying to do Metric Learning with a Contrastive Loss, but my CNN is not learning because ...
- 11
0
votes
0
answers
88
views
Value accuracy remains the same
I have used my own build model and also fine-tuned other two model ResNeT50 and VGG16, but val_acc remains the same for them all.
...
- 21
0
votes
1
answer
218
views
False positive in Multi class Image classification
I am training a neural network with some convolution layers for multi class image classification. I am using keras to build and train the model. I am using 1600 images for all categories for training. ...
- 1
0
votes
1
answer
47
views
Discrepancy between cross-validation and un-seen data predictions
I am facing an issue with an imbalanced dataset. The dataset contains 20% targets and 80% non-targets. I am expecting a confusion matrix below when I give un-seen data to the trained model.
...
0
votes
1
answer
26
views
Can we not backpropagate model
I saw a model based on CNN for question classification. The author said that they don't backpropagate gradient to embeddings. How this is possible to update network if you don't backpropagate please?
...
- 231
-1
votes
1
answer
128
views
How to show combined overall accuracy for a multi-ouput model in Keras?
I have a model of the following structure. It has 6 outputs. Given an image, the model predicts classes of 6 different components from the image.
The metrics I used are:
As you can see it outputs an ...
- 1
2
votes
1
answer
633
views
Input shape of dataset in hybrid CNN-SVM classifier
I am working on hybrid CNN-SVM for classification task, where I aim to use CNN for feature extraction and SVM for classification.
So after the training of my CNN model as below:
...
- 125
0
votes
0
answers
71
views
Why CNN Linear Regression predicting always same value?
I have dataset with around 3 million samples which almost fit to gauss distribution. X-axis are
normalized target values.
I am using WRN model and if i am solving binary or multi-class classification ...
- 1
0
votes
1
answer
61
views
How pixel value range is satisfying in CNN
gray-scale images have pixel value range [0,1]. In most imaging task such as Denoising, deblurring, and inpainting we usually calculate mean square error of observed image and denoised image. However, ...
- 21
2
votes
1
answer
350
views
RCNN to predict sequence of images (video frames)?
In the following work the authors apply a convolutional recurrent neural network (RNN) to predict the spatiotemporal evolution of microstructure represented by 2D image sequences. In particular, they ...
- 123
0
votes
0
answers
131
views
The best way to work with hybrid CNN-SVM
I am working on a hybrid CNN-SVM where I aim to use CNN for feature extraction and SVM for classification. However, I am confused as after reading related works, I found many approaches:
-Some use SVM ...
- 125
0
votes
0
answers
23
views
How to train a CNN with several features but test with less features
I'm trying to train a model (with TensorFlow and Keras) to classify the soil based on the x,y, and z coordinates. I have a table 8321 x 10 where 8321 are the points in a mesh, and 10 are the features ...
- 101
1
vote
1
answer
51
views
Advice / Good practises | CNN poor image diversity
I am currenty working on a project that involves multiple cameras fixed on the ceiling. Each time I take a picture, I check whether there is a "cart" right under the camera.
I would like to ...
- 11
0
votes
0
answers
24
views
Inbetween CNN and MLP: neural network architecture for "close to convolutional" problem?
I am looking to approximate an (expensive to calculate precisely) forward problem using a NN. Input and output are vectors of identical length. Although not linear, the output somewhat resembles a ...
- 1
1
vote
0
answers
175
views
CNN auto-encoder performs much worse than Fully connected auto-encoder
I am trying to develop a one-class classifier, that will learn regular examples, and, hopefully, will have hard times reconstructing anomaly observations.
I have 1D signals which I tried to ...
- 151