Julian de Wit

Hello, Atm I made a very wild "practise download" for a kaggle competion I'm doing. I made a big mess of it but I learned a lot about the codebase....

My own homebrew stuff..

Hello.. you can count the pixel intensities. Perhaps with a certain threshold. This is also done to determine the final volume. (see count_pixels)

At the time of building I believe keras had no deconvolution but.. Actually recently a u-net was published. https://github.com/jocicmarko/ultrasound-nerve-segmentation I used mxnet because of better multi-platform support. Theano cannot be...

Hello, thanks. First of all the dropout part is still something under heavy debate for FCN as I understand. 4 or 5 made no real big difference although I agree...

It helped a little but not much if I remember correctly.

Hello thanks, I decided to assume my hand-labeled traindata as "given". Like in MNIST the dataset is also given, Was there anything you would like to do with the files...

Hello, 1. I just do (logistic) regression for every pixel so I just need one channel.. Perhaps they use softmax for every pixel ? Softmax needs 2 values per pixel....