You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm pretty new to Machine Learning/Deep Learning. For my bachelor thesis I am trying to recreate this Paper with other Neural Networks.
(Basically a classifier behind a fine tuned Autoencoder)
I know that Unet is meant to segment images. But I'm trying to reconstruct images in RGB with this Network.
I tried to feed the network RGB masks (which where just the original images) but I don't know how to modify the DataProvider to take output images with 3 channels.
Furthermore I would try to nullify the concat_and_crop connections, to prevent the network to just take the original image.
Any suggestions or better alternatives for this project?
The text was updated successfully, but these errors were encountered:
Hey folks,
I'm pretty new to Machine Learning/Deep Learning. For my bachelor thesis I am trying to recreate this Paper with other Neural Networks.
(Basically a classifier behind a fine tuned Autoencoder)
I know that Unet is meant to segment images. But I'm trying to reconstruct images in RGB with this Network.
I tried to feed the network RGB masks (which where just the original images) but I don't know how to modify the DataProvider to take output images with 3 channels.
Furthermore I would try to nullify the concat_and_crop connections, to prevent the network to just take the original image.
Any suggestions or better alternatives for this project?
The text was updated successfully, but these errors were encountered: