I am trying to train a feed forward neural net for image classification. In this process, I am implementing a restricted boltzmann machine to help pre-train the weights.

My images are grey-scale with pixel intensities from 0-255.

I was wondering how to sample and turn these pixels into binary since their intensities do not offer a representation of their probabilities. I also normalized these features to zero mean and unit variance.


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.