By default using DeepDream technique you can creating a dreamlike image out of two different images.

Is it possible to easily enhance this technique to generate one image out from three?

  • 1
    $\begingroup$ Absolutely. The convolutional neural network behind DeepDream can be trained using how many input/sample images you wish. More patterns in the training set determine a more pareidolic output. $\endgroup$ – Lovecraft Jan 13 '17 at 23:16
  • $\begingroup$ @Lovecraft please put this as an answer. $\endgroup$ – FreezePhoenix Apr 14 '18 at 17:48

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.