evgevd's profile - activity

np.zeros((20,64,64,3)) as input gives array([[0.99938166]], dtype=float32) as output Keras model in .h5, .pb and .pbtxt

np.zeros((20,64,64,3)) as input gives array([[0.99938166]], dtype=float32) as output Keras model in .h5, .pb and .pbtxt

np.zeros((20,64,64,3)) as input gives array([[0.99938166]], dtype=float32) as output Keras model in .h5, .pb and .pbtxt

https://answers.opencv.org/question/191013/dnn-loading-of-models-and-order-of-channels/ So it did not become clearer.

I tried. But it doesnt work. I see that my original keras model uses «channels last» so input array has shape (20, 64,

The same result on python with cv__version__ = 4.2.0 Did I correctly understand? You got multiple output after blobFrom

It is a binary classificator. When I run it on single images i got correct outputs. cv_pred = [] for img in imgs:

It is a binary classificator. When I run it on single images i got correct outputs. cv_pred = [] for img in imgs:

OpenCV 4.1. I use CNN with 64x64x3 on input and sigmoid on output layer. I got it by converting keras→tensorflow→openCV

OpenCV 4.1. I use CNN with 64x64x3 on input sigmoid on output layer. I got it by converting keras→tensorflow→openCV DNN.

DNN forwards only first image I make input for neural network with cv.dnn.blobFromImages, and then run forward function

DNN forwards only first image I make input for neural network with cv.dnn.blobFromImages, and then run forward function

DNN forwards only first image I make input for neural network with cv.dnn.blobFromImages, and then run forward function

I tried with normalization and got the same result: from sklearn.preprocessing import normalize train_features = normal

I tried with normalization and got the same result: from sklearn.preprocessing import normalize train_features = normal

I tried with normalization and got the same result: from sklearn.preprocessing import normalize train_features = norm

train_features.shape ((448, 4) train_labels.shape (448,))

train_features.shape, train_labels.shape ((448, 4), (448,))

No I didnt. Is it necessary in this case?

Logistic Regression doesnt learn anything I train binary LR in python (opencv 3.1 and 4.1) and get always prediction = 1

It works! Thanx!

I tried to perform training on my data and on toy example too. Command prompt doesnt show anything but the message about restarting.

segm_features = np.array([[1,0], [1,1], [0,1]])
labels = np.array([1,0,0])
cv_model = cv.ml.DTrees_create()
dtype = np.int32
train_data = cv.ml.TrainData_create(segm_features.astype(np.float32), cv.ml.ROW_SAMPLE, labels.astype(dtype))

cv_model.train(train_data)

3.4.2.17 with 4.1.0.25 too

Jupyter kernel crashes after running "train" for decision tree classifier I tried to perform training on my data and on