OpenCV - Trainings lead to different result when using TrainData_create
0
I'm using MLP ANN provided by OpenCV 3.4 with python. I noticed that when training data is prepared via cv2.ml.TrainData_create the ANN performs well, in case this is not used but same samples and parameters are used, ANN is not correctly trained.
Here I don't mean trainings differ even if using the same data (which can be expected due to random starting points), because what I see here is working-training VS not-working-training and this occurs always.
The following code uses cv2.ml.TrainData_create
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
td = cv2.ml.TrainData_create(input_array, cv2.ml.ROW_SAMPLE, output_array)
ann.train(td, cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(td, cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
and this works well:
print(ann.predict(input_array))
(0.0, array([[ 1.0000000e+00, 0.0000000e+00, 4.7625793e-16, 2.8575474e-16],
[ 1.9050316e-16, 10000000e+00, 5.7150949e-16, 9.5251581e-17],
[ 9.5251581e-17, 0.0000000e+00, 1.0000000e+00, -1.9050316e-16],
[-1.9050316e-16, -1.9050316e-16, 0.0000000e+00, 1.0000000e+00]],
dtype=float32))
The following code doesn't use cv2.ml.TrainData_create but apparently use the same data and parameters:
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP | cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(np.array([[1.0, 0.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[1.0, 0.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 1.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 1.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 0.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 1.0, 0.0]], dtype=np.float32))
ann.train(np.array([[1.0, 1.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 0.0, 1.0]], dtype=np.float32))
But this one simply doesn't work:
print(ann.predict(input_array))
(0.0, array([[ 1.2886142 , 0.51306236, -1.0352006 , -0.19007786],
[ 1.2194023 , 0.7686653 , -1.097198 , -0.03246666],
[ 0.99483347, 0.40380374, -0.917998 , 0.08949649],
[ 0.7475754 , 0.12770385, -0.81321925, 0.37416443]],
dtype=float32))
What's wrong with the second code snippet?
python opencv machine-learning neural-network training-data
asked Nov 23 '18 at 8:21
kumakuma
12
add a comment |
0
I'm using MLP ANN provided by OpenCV 3.4 with python. I noticed that when training data is prepared via cv2.ml.TrainData_create the ANN performs well, in case this is not used but same samples and parameters are used, ANN is not correctly trained.
Here I don't mean trainings differ even if using the same data (which can be expected due to random starting points), because what I see here is working-training VS not-working-training and this occurs always.
The following code uses cv2.ml.TrainData_create
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
td = cv2.ml.TrainData_create(input_array, cv2.ml.ROW_SAMPLE, output_array)
ann.train(td, cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(td, cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
and this works well:
print(ann.predict(input_array))
(0.0, array([[ 1.0000000e+00, 0.0000000e+00, 4.7625793e-16, 2.8575474e-16],
[ 1.9050316e-16, 10000000e+00, 5.7150949e-16, 9.5251581e-17],
[ 9.5251581e-17, 0.0000000e+00, 1.0000000e+00, -1.9050316e-16],
[-1.9050316e-16, -1.9050316e-16, 0.0000000e+00, 1.0000000e+00]],
dtype=float32))
The following code doesn't use cv2.ml.TrainData_create but apparently use the same data and parameters:
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP | cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(np.array([[1.0, 0.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[1.0, 0.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 1.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 1.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 0.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 1.0, 0.0]], dtype=np.float32))
ann.train(np.array([[1.0, 1.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 0.0, 1.0]], dtype=np.float32))
But this one simply doesn't work:
print(ann.predict(input_array))
(0.0, array([[ 1.2886142 , 0.51306236, -1.0352006 , -0.19007786],
[ 1.2194023 , 0.7686653 , -1.097198 , -0.03246666],
[ 0.99483347, 0.40380374, -0.917998 , 0.08949649],
[ 0.7475754 , 0.12770385, -0.81321925, 0.37416443]],
dtype=float32))
What's wrong with the second code snippet?
python opencv machine-learning neural-network training-data
asked Nov 23 '18 at 8:21
kumakuma
12
add a comment |
0
0
0
I'm using MLP ANN provided by OpenCV 3.4 with python. I noticed that when training data is prepared via cv2.ml.TrainData_create the ANN performs well, in case this is not used but same samples and parameters are used, ANN is not correctly trained.
Here I don't mean trainings differ even if using the same data (which can be expected due to random starting points), because what I see here is working-training VS not-working-training and this occurs always.
The following code uses cv2.ml.TrainData_create
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
td = cv2.ml.TrainData_create(input_array, cv2.ml.ROW_SAMPLE, output_array)
ann.train(td, cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(td, cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
and this works well:
print(ann.predict(input_array))
(0.0, array([[ 1.0000000e+00, 0.0000000e+00, 4.7625793e-16, 2.8575474e-16],
[ 1.9050316e-16, 10000000e+00, 5.7150949e-16, 9.5251581e-17],
[ 9.5251581e-17, 0.0000000e+00, 1.0000000e+00, -1.9050316e-16],
[-1.9050316e-16, -1.9050316e-16, 0.0000000e+00, 1.0000000e+00]],
dtype=float32))
The following code doesn't use cv2.ml.TrainData_create but apparently use the same data and parameters:
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP | cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(np.array([[1.0, 0.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[1.0, 0.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 1.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 1.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 0.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 1.0, 0.0]], dtype=np.float32))
ann.train(np.array([[1.0, 1.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 0.0, 1.0]], dtype=np.float32))
But this one simply doesn't work:
print(ann.predict(input_array))
(0.0, array([[ 1.2886142 , 0.51306236, -1.0352006 , -0.19007786],
[ 1.2194023 , 0.7686653 , -1.097198 , -0.03246666],
[ 0.99483347, 0.40380374, -0.917998 , 0.08949649],
[ 0.7475754 , 0.12770385, -0.81321925, 0.37416443]],
dtype=float32))
What's wrong with the second code snippet?
python opencv machine-learning neural-network training-data
asked Nov 23 '18 at 8:21
kumakuma
12
I'm using MLP ANN provided by OpenCV 3.4 with python. I noticed that when training data is prepared via cv2.ml.TrainData_create the ANN performs well, in case this is not used but same samples and parameters are used, ANN is not correctly trained.
Here I don't mean trainings differ even if using the same data (which can be expected due to random starting points), because what I see here is working-training VS not-working-training and this occurs always.
The following code uses cv2.ml.TrainData_create
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
td = cv2.ml.TrainData_create(input_array, cv2.ml.ROW_SAMPLE, output_array)
ann.train(td, cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(td, cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
and this works well:
print(ann.predict(input_array))
(0.0, array([[ 1.0000000e+00, 0.0000000e+00, 4.7625793e-16, 2.8575474e-16],
[ 1.9050316e-16, 10000000e+00, 5.7150949e-16, 9.5251581e-17],
[ 9.5251581e-17, 0.0000000e+00, 1.0000000e+00, -1.9050316e-16],
[-1.9050316e-16, -1.9050316e-16, 0.0000000e+00, 1.0000000e+00]],
dtype=float32))
The following code doesn't use cv2.ml.TrainData_create but apparently use the same data and parameters:
import cv2
import numpy as np
ann = cv2.ml.ANN_MLP_create()
ann.setTrainMethod(cv2.ml.ANN_MLP_BACKPROP | cv2.ml.ANN_MLP_UPDATE_WEIGHTS | cv2.ml.ANN_MLP_NO_INPUT_SCALE | cv2.ml.ANN_MLP_NO_OUTPUT_SCALE)
ann.setActivationFunction(cv2.ml.ANN_MLP_SIGMOID_SYM)
ann.setLayerSizes(np.array([3, 8, 4]))
ann.setTermCriteria(( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 ))
input_array = np.array([ [1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0],
[1.0, 1.0, 1.0],
], dtype=np.float32)
output_array = np.array([ [1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0],
], dtype=np.float32)
SAMPLES = 5000
for x in range(0, SAMPLES):
ann.train(np.array([[1.0, 0.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[1.0, 0.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 1.0, 0.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 1.0, 0.0, 0.0]], dtype=np.float32))
ann.train(np.array([[0.0, 0.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 1.0, 0.0]], dtype=np.float32))
ann.train(np.array([[1.0, 1.0, 1.0]], dtype=np.float32), cv2.ml.ROW_SAMPLE, np.array([[0.0, 0.0, 0.0, 1.0]], dtype=np.float32))
But this one simply doesn't work:
print(ann.predict(input_array))
(0.0, array([[ 1.2886142 , 0.51306236, -1.0352006 , -0.19007786],
[ 1.2194023 , 0.7686653 , -1.097198 , -0.03246666],
[ 0.99483347, 0.40380374, -0.917998 , 0.08949649],
[ 0.7475754 , 0.12770385, -0.81321925, 0.37416443]],
dtype=float32))
What's wrong with the second code snippet?
python opencv machine-learning neural-network training-data
python opencv machine-learning neural-network training-data
asked Nov 23 '18 at 8:21
kumakuma
12
asked Nov 23 '18 at 8:21
kumakuma
12
asked Nov 23 '18 at 8:21
kumakuma
12
asked Nov 23 '18 at 8:21
kumakuma
12
asked Nov 23 '18 at 8:21
kumakuma
12
12
add a comment |
add a comment |
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%2f53442938%2fopencv-trainings-lead-to-different-result-when-using-traindata-create%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%2f53442938%2fopencv-trainings-lead-to-different-result-when-using-traindata-create%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
