How do I obtain filters from convolutional neural network(CNN)? My idea is something like this: do random images of the input images (28x28) and get random patches (8x8). Then use autoencoders to learn the common features of the patches (features = hidden units; approximately 100, for example). Then apply features filters to the input images and do convolution. Am I correct?

I am confused because sometime the literature state only using like, e.g. 8, filters, but in my case I have 100.

  • $\begingroup$ What's the CNN's output? I.e. what's the learning task? $\endgroup$ – Franck Dernoncourt Jan 3 '16 at 16:30
  • $\begingroup$ @FranckDernoncourt let say the output is classifying 10 digits. $\endgroup$ – RockTheStar Jan 5 '16 at 2:51

Convolutional filters are usually trained along with the rest of the network weights. The Tensorflow tutorial is a good place to start understanding convolutional neural networks in practice. https://www.tensorflow.org/versions/r1.0/get_started/mnist/beginners


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.