Why feature vectors have a lot of zero values in Keras VGG16 model's output?
.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty{ height:90px;width:728px;box-sizing:border-box;
}
0
I am trying to extract the features from the last layer of VGG16 model in Keras using the following code:
from keras.applications.vgg16 import VGG16
from keras.preprocessing import image
from keras.applications.vgg16 import preprocess_input
import numpy as np
model = VGG16(weights='imagenet', include_top=True )
img_path = 'E:projectKERAS DEEPpoodle.png'
img = image.load_img(img_path, target_size=(224, 224))
img_data = image.img_to_array(img)
img_data = np.expand_dims(img_data, axis=0)
img_data = preprocess_input(img_data)
model.summary()
model.layers.pop();
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes =
feature = model.predict(img_data)[0]
feature variable supposes to be the feature vector but it has a lot of zeros, I think because of relu layer. In Matlab, for example, the extracted features vector seems to have both positive and negative values, how I can get the same with keras model?
The matlab code is:
im=imread('poodle.png');
im=imresize(im,[224,224]);
net=vgg16;
trainingFeatures = activations(net, im, 'fc7', ...
'OutputAs', 'rows');
the two output vectors feature and trainingFeatures as following (python output to the left and Matlab's to the right
And here is the tested image:
python tensorflow keras deep-learning feature-extraction
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
|
show 7 more comments
0
I am trying to extract the features from the last layer of VGG16 model in Keras using the following code:
from keras.applications.vgg16 import VGG16
from keras.preprocessing import image
from keras.applications.vgg16 import preprocess_input
import numpy as np
model = VGG16(weights='imagenet', include_top=True )
img_path = 'E:projectKERAS DEEPpoodle.png'
img = image.load_img(img_path, target_size=(224, 224))
img_data = image.img_to_array(img)
img_data = np.expand_dims(img_data, axis=0)
img_data = preprocess_input(img_data)
model.summary()
model.layers.pop();
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes =
feature = model.predict(img_data)[0]
feature variable supposes to be the feature vector but it has a lot of zeros, I think because of relu layer. In Matlab, for example, the extracted features vector seems to have both positive and negative values, how I can get the same with keras model?
The matlab code is:
im=imread('poodle.png');
im=imresize(im,[224,224]);
net=vgg16;
trainingFeatures = activations(net, im, 'fc7', ...
'OutputAs', 'rows');
the two output vectors feature and trainingFeatures as following (python output to the left and Matlab's to the right
And here is the tested image:
python tensorflow keras deep-learning feature-extraction
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
1
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Which variables are you showing in the two pictures?featureandtrainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?
– Dan
Nov 22 '18 at 12:11
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
1
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33
|
show 7 more comments
0
0
0
I am trying to extract the features from the last layer of VGG16 model in Keras using the following code:
from keras.applications.vgg16 import VGG16
from keras.preprocessing import image
from keras.applications.vgg16 import preprocess_input
import numpy as np
model = VGG16(weights='imagenet', include_top=True )
img_path = 'E:projectKERAS DEEPpoodle.png'
img = image.load_img(img_path, target_size=(224, 224))
img_data = image.img_to_array(img)
img_data = np.expand_dims(img_data, axis=0)
img_data = preprocess_input(img_data)
model.summary()
model.layers.pop();
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes =
feature = model.predict(img_data)[0]
feature variable supposes to be the feature vector but it has a lot of zeros, I think because of relu layer. In Matlab, for example, the extracted features vector seems to have both positive and negative values, how I can get the same with keras model?
The matlab code is:
im=imread('poodle.png');
im=imresize(im,[224,224]);
net=vgg16;
trainingFeatures = activations(net, im, 'fc7', ...
'OutputAs', 'rows');
the two output vectors feature and trainingFeatures as following (python output to the left and Matlab's to the right
And here is the tested image:
python tensorflow keras deep-learning feature-extraction
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
I am trying to extract the features from the last layer of VGG16 model in Keras using the following code:
from keras.applications.vgg16 import VGG16
from keras.preprocessing import image
from keras.applications.vgg16 import preprocess_input
import numpy as np
model = VGG16(weights='imagenet', include_top=True )
img_path = 'E:projectKERAS DEEPpoodle.png'
img = image.load_img(img_path, target_size=(224, 224))
img_data = image.img_to_array(img)
img_data = np.expand_dims(img_data, axis=0)
img_data = preprocess_input(img_data)
model.summary()
model.layers.pop();
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes =
feature = model.predict(img_data)[0]
feature variable supposes to be the feature vector but it has a lot of zeros, I think because of relu layer. In Matlab, for example, the extracted features vector seems to have both positive and negative values, how I can get the same with keras model?
The matlab code is:
im=imread('poodle.png');
im=imresize(im,[224,224]);
net=vgg16;
trainingFeatures = activations(net, im, 'fc7', ...
'OutputAs', 'rows');
the two output vectors feature and trainingFeatures as following (python output to the left and Matlab's to the right
And here is the tested image:
python tensorflow keras deep-learning feature-extraction
python tensorflow keras deep-learning feature-extraction
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
edited Nov 22 '18 at 12:52
Ahmed Tarawneh
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
asked Nov 21 '18 at 20:54
Ahmed TarawnehAhmed Tarawneh
167
167
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
1
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Which variables are you showing in the two pictures?featureandtrainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?
– Dan
Nov 22 '18 at 12:11
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
1
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33
|
show 7 more comments
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
1
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Which variables are you showing in the two pictures?featureandtrainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?
– Dan
Nov 22 '18 at 12:11
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
1
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
1
1
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Which variables are you showing in the two pictures?
feature and trainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?– Dan
Nov 22 '18 at 12:11
Which variables are you showing in the two pictures?
feature and trainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?– Dan
Nov 22 '18 at 12:11
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
1
1
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33
|
show 7 more comments
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
draft saved
draft discarded
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53420338%2fwhy-feature-vectors-have-a-lot-of-zero-values-in-keras-vgg16-models-output%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
draft saved
draft discarded
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
draft saved
draft discarded
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53420338%2fwhy-feature-vectors-have-a-lot-of-zero-values-in-keras-vgg16-models-output%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
There is no way to answer this if you don't include the equivalent Matlab code and the results you are seeing.
– Matias Valdenegro
Nov 21 '18 at 22:50
1
Edited! I think its more clear now :)
– Ahmed Tarawneh
Nov 22 '18 at 12:05
Which variables are you showing in the two pictures?
featureandtrainingFeatures? Also where is ResNet coming into this, it looks like you've only used VGG? Is your python model using ReLu and your MATLAB model using tanh maybe?– Dan
Nov 22 '18 at 12:11
Yes, feature and trainingFeatures. I have edited the code and used vgg16 in both Python and Matlab for more clarification. Keras and Matlab model (vgg16) use the same ReLu layer
– Ahmed Tarawneh
Nov 22 '18 at 12:17
1
Yes, it seems to be before ReLU, you can't extract the same in Keras as each layer has the ReLU embedded inside the layer. This explains all the differences.
– Matias Valdenegro
Nov 22 '18 at 13:33