All Questions
13 questions
1
vote
0
answers
61
views
Mirror model to create autoencoder. (Keras)
I have a model in keras and I want to use it as a feature extractor in an autoencoder configuration. The first test I want to do is to mirror the model and use the original version as the encoder, and ...
- 33
1
vote
1
answer
818
views
CNN autoencoder latent space representation meaning
I am using a CNN autoencoder to denoise some syntetic noisy data I have generated. The purpouse of this exercise is to test the denoising capabilities of denoising autoencoder.
A part this I also ...
- 168
1
vote
1
answer
177
views
Convolutional Autoencoder Not Training on (62,47,1) dataset, "Expected Shape Error"
I'm trying to implement a Convolutional Autoencoder on the Faces in The Wild Dataset, which is made up of images shaped 62x47x3.
However, the Keras Convolutional Autoencoder example on the mnist ...
- 47
4
votes
1
answer
3k
views
High loss from convolutional autoencoder keras
I am training a convolutional autoencoder and I am having trouble getting the loss to decrease and was hoping someone could point out some possible improvements.
I have greyscale images of 1024x1024 (...
0
votes
0
answers
277
views
Keras Convolution layer kernel size with more than 3 channels
I have a question regarding the filter of a convolutional layer in keras. I have image data with 4 channels which are fed through a 3D convolutional network, and Iam not sure how to set up the Conv3D ...
0
votes
1
answer
396
views
Why does this Autoencoder consisting of all convolutions keep pushing the output to a blank/white image?
I am having a lot of trouble understanding the behaviour of my model and need some help to try figure it out.
Suppose this Autoencoder consisting of all convolution layers:
initializer = he_uniform()
...
- 876
0
votes
1
answer
3k
views
Why doesn't the UpSampling2d Keras layer work?
I tried to build a convolutional autoencoder in keras but it doesn't seem to work properly.
First of all, here's the Code:
from keras.models import Sequential
from keras.layers import Reshape
from ...
- 297
5
votes
1
answer
4k
views
Can I use rectangle images with a convolution neural network in Keras?
Say I'd like to use Keras's Convolutional2D function to build a CNN, can the input image be of size [224, 320, 3] instead of something like [224, 224, 3]?
Should I keep my images in their rectangle ...
- 406
2
votes
1
answer
4k
views
Conv1D convolutional Autoencoder for text in keras
How I solve this issue
ValueError: Error when checking target: expected conv-decode3 to have shape (None, 14, 300) but got array with shape (6559, 16, 300)
NUM_WORDS = 3
pool_size = 2
x = Input(shape=...
- 159
4
votes
1
answer
18k
views
How to implement a 1D Convolutional Auto-encoder in Keras for vector data?
My input vector to the auto-encoder is of size 128. I have 730 samples in total (730x128). I am trying to use a 1D CNN auto-encoder. I would like to use the hidden layer as my new lower dimensional ...
1
vote
1
answer
2k
views
Variational Autoencoder loss is increasing
This is my particular code for creating an abnormal convolutional autoencoder and my problem is the loss function is not able to converge to anything at all. I have tried different optimizers for ...
- 61
3
votes
1
answer
655
views
Cropping in the very last layer in autoencoder in keras
I have images of shape 391 x 400. I attempted to use the autoencoder as described here.
Specifically, I have used the following code:
from keras.layers import Input, Dense, Conv2D, MaxPooling2D, ...
- 1,189
6
votes
1
answer
3k
views
Fully convolutional autoencoder for variable-sized images in keras
I want to build a convolutional autoencoder where the size of the input in not constant. I'm doing that by stacking up conv-pool layers until I reach an encoding layer, and then doing the reverse with ...
- 489