Feeding separate image "channels" into separate paths of a CNN

[gotk...@gmail.com]

Hi all,

I'm working on image segmentation using 3D images. I'd like to implement one of my CNNs in Keras. There is only one color channel for each image. For each voxel, I extract a patch in the xy-, yz-, and xz-planes. I want each of these patches to go into their own paths in the CNN, until the final layers where I concatenate the feature map outputs and train a softmax classifier to classify the voxel.

Does anyone know how to do this? The only thing I can think of right now is training 3 CNNs separately, then removing the top layers from each CNN and train a final softmax classifier on the resulting output... however this seems inefficient as I would have to generate many times more feature maps for each patch (there are millions of patches). 

Thanks,

Ian

[Daπid]

You can make a multi-input model just like this:

https://keras.io/getting-started/functional-api-guide/#multi-input-and-multi-output-models

You have several input branches that are concatenated at the end.

If you want to share weights across branches, you can follow the shared vision example:

https://keras.io/getting-started/functional-api-guide/#more-examples

--
You received this message because you are subscribed to the Google Groups "Keras-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keras-users+unsubscribe@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/54950981-4029-409b-8a6f-dc9c44206566%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.