Basic concept for feature extraction, used in Convolutional Neural Netwrok
Create image with numpy
Perform Convolution Operation with kernel size-->3*3 and stride=1
If used convolve2d--->Note that the kernal has to be reversed. Otherwise usesignal.correlate2d.
Perfrom pooling operation(in this case maxpool) with size--->2*2 and stride=2
Make a single vector for futher processing
Pass Vector to artificial neural network..... Thanks for Viewing this notebook. Hope this provide a basic concept for feature extraction, used in Convolutional Neural Network