please write requirements.txt as many tensorflow functions are deprecated and will be removed in future versions. by this file we can make a virtual environment for the project so that it runs well.
please tell me if there is a way to extract feature of these signature pic as this is really huge data.

For how many epoch I should run the model

I use file SigNet-BHSig260.ipynb on Google Colab, but when i ran code:

network definition

base_network = create_base_network_signet(input_shape)

input_a = Input(shape=(input_shape))
input_b = Input(shape=(input_shape))

because we re-use the same instance base_network,

the weights of the network

will be shared across the two branches

processed_a = base_network(input_a)
processed_b = base_network(input_b)

Compute the Euclidean distance between the two vectors in the latent space

distance = Lambda(euclidean_distance, output_shape=eucl_dist_output_shape)([processed_a, processed_b])

model = Model(input=[input_a, input_b], output=distance)

==>

TypeError Traceback (most recent call last)
in ()
1 # network definition
----> 2 base_network = create_base_network_signet(input_shape)
3
4 input_a = Input(shape=(input_shape))
5 input_b = Input(shape=(input_shape))

5 frames
/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/utils/generic_utils.py in validate_kwargs(kwargs, allowed_kwargs, error_message)
806 for kwarg in kwargs:
807 if kwarg not in allowed_kwargs:
--> 808 raise TypeError(error_message, kwarg)
809
810

TypeError: ('Keyword argument not understood:', 'init')
###############################################################################
And this is ''create_base_network_signet'' fuction:

def create_base_network_signet(input_shape):
'''Base Siamese Network'''

seq = Sequential()
seq.add(Conv2D(96, kernel_size=(11, 11), activation='relu', name='conv1_1', strides=4, input_shape= input_shape, 
                    init='glorot_uniform', dim_ordering='tf'))
seq.add(BatchNormalization(epsilon=1e-06, mode=0, axis=1, momentum=0.9))
seq.add(MaxPooling2D((3,3), strides=(2, 2)))    
seq.add(ZeroPadding2D((2, 2), dim_ordering='tf'))

seq.add(Conv2D(256, kernel_size=(5, 5), activation='relu', name='conv2_1', strides=1, init='glorot_uniform',  dim_ordering='tf'))
seq.add(BatchNormalization(epsilon=1e-06, mode=0, axis=1, momentum=0.9))
seq.add(MaxPooling2D((3,3), strides=(2, 2)))
seq.add(Dropout(0.3))# added extra
seq.add(ZeroPadding2D((1, 1), dim_ordering='tf'))

seq.add(Conv2D(384, kernel_size=(3, 3), activation='relu', name='conv3_1', strides=1, init='glorot_uniform',  dim_ordering='tf'))
seq.add(ZeroPadding2D((1, 1), dim_ordering='tf'))

seq.add(Conv2D(256, kernel_size=(3, 3), activation='relu', name='conv3_2', strides=1, init='glorot_uniform', dim_ordering='tf'))    
seq.add(MaxPooling2D((3,3), strides=(2, 2)))
seq.add(Dropout(0.3))# added extra
seq.add(Flatten(name='flatten'))
seq.add(Dense(1024, W_regularizer=l2(0.0005), activation='relu', init='glorot_uniform'))
seq.add(Dropout(0.5))

seq.add(Dense(128, W_regularizer=l2(0.0005), activation='relu', init='glorot_uniform')) # softmax changed to relu

return seq

###########################################################################

How can I fix this error? Thank you so much!

Hello hlamba28,

I have trained the code on 2 dataset CEDAR and BHSig260. I have created train, validation and test set with following partition

There are total 276 (genuine - genuine) signature pairs and 276 (genuine - forged) signature pairs per users for all the dataset.

I am getting results as shown below

Why accuracy on Bengali and Hindi signatures is low ?

You claimed Accuracy = 81.71% and Threshold = 0.55 for Hindi Signatures, but I am getting only 50% I tried changing train, validation and test size but test accuracy is stuck 50% not less not more

hello. i am just curious if the same code will work well if i happen to use gpu~

I need help in plotting FAR or FRR