0
$\begingroup$

I have a network which has a input size of (28x28x1) and since I'm using (3x3 convolution) so the receptive field is (3x3). Before going further I will show the code snippet

from keras.layers import Activation, MaxPooling2D

model = Sequential()# adding sequential layer
model.add(Convolution2D(32, 3, 3, activation='relu', input_shape=(28,28,1)))# receptive field = (3X3) input channel dimension = (28x28x1)
model.add(Convolution2D(64, 3, 3, activation='relu'))# receptive field = (5X5) input channel dimension = (26x26x32)
model.add(Convolution2D(128, 3, 3, activation='relu'))# receptive field = (7X7) input channel dimension = (24x24x64)

model.add(MaxPooling2D(pool_size=(2, 2)))# receptive field = (7X7) input channel dimension = (22x22x128)

model.add(Convolution2D(256, 3, 3, activation='relu'))# receptive field = (9X9) input channel dimension = (11x11x128)
model.add(Convolution2D(512, 3, 3, activation='relu'))# receptive field = (11x11) input channel dimension = (9x9x256)
model.add(Convolution2D(1024, 3, 3, activation='relu'))# receptive field = (13x13) input channel dimension = (7x7x512)
model.add(Convolution2D(2048, 3, 3, activation='relu'))# receptive field = (15x15) input channel dimension = (5x5x1024)
model.add(Convolution2D(10, 3, 3, activation='relu'))# receptive field = (17x17) input channel dimension = (3x3x2048)

model.add(Flatten())# receptive field = (17x17) input channel dimension = (1x1x10)
model.add(Activation('softmax'))# adding activation layer

model.summary()# will give the whole summary of the network 

In the network before the flattening of the network the receptive field = (17x17) and the input channel dimension= (1x1x10).

My question is will the network perform worse when the receptive field becomes larger than the input image(input channel dimension) ?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.