@lidaryani I got it working by building Jonlong's caffe version and running Cascaded-FCN with that one. See also the Cascaded-FCN jupyter notebook and issue #3

from cascaded-fcn.

@wwlaoxi Hi, have you solved this problems ?

from cascaded-fcn.

Same problem here. I'm running the notebook in an official Caffe release docker in CPU mode.

@mohamed-ezz @PatrickChrist @FelixGruen Any help/hints towards solving this would be appreciated!

from cascaded-fcn.

I also achieved the same result. The predicted slice was not a satisfactory result. With regard to the previous issue (#31 ), I tested it in CPU mode and was unacceptable again.
@mohamed-ezz @PatrickChrist @FelixGruen Do we have to consider a particular item in the implementation?

from cascaded-fcn.

Hello @lidaryani , please see this note about the Caffe version and the crop layer #3 (comment)

from cascaded-fcn.

I used the docker image (Jon Long) on new CT volumes.
The inference was impressive.
Using Caffe 1.0.0, I am getting identical results to the above Prediction screenshot.

Could somebody please provide an example of updating the model r/e the crop layer--
#3 (comment)

from cascaded-fcn.

I used the docker image (Jon Long) on new CT volumes.
The inference was impressive.
Using Caffe 1.0.0, I am getting identical results to the above Prediction screenshot.

Could somebody please provide an example of updating the model r/e the crop layer--
#3 (comment)

Here is my updated crop layers which appeared to work under Cafe 1.0.0 (AWS Python 3 configured)

layer {
name: "crop_d3c-d3cc"
type: "Crop"
bottom: "d3c" *current blob size (1, 512, 64, 64)
bottom: "u3a" *desired blob size (1, 512, 56, 56)
top: "d3cc"
crop_param {
axis: 2
offset: 4
offset: 4
}
}
layer {

type: "Crop"
bottom: "d2c" (1, 256, 136, 136)
bottom: "u2a" *desired blob size (1, 256, 104, 104)
top: "d2cc"
crop_param {
axis: 2
offset: 16
offset: 16
}
}
layer {
name: "crop_d1c-d1cc"
type: "Crop"
bottom: "d1c" (1, 128, 280, 280)
bottom: "u1a" *desired blob size (1, 128, 200, 200)
top: "d1cc"
crop_param {
axis: 2
offset: 40
offset: 40
}
}
layer {
name: "crop_d0c-d0cc"
type: "Crop"
bottom: "d0c" (1, 64, 568, 568)
bottom: "u0a" *desired blob size (1, 64, 392, 392)
top: "d0cc"
crop_param {
axis: 2
offset: 88
offset: 88
}
}

offset = (current blob size - desired blob size) / 2

I am down to just the following warning which I hope are related to training and not inference:
I0610 23:47:14.142359 4565 net.cpp:744] Ignoring source layer bn_d0b (batch normalization?)
I0610 23:47:14.169219 4565 net.cpp:744] Ignoring source layer loss (loss for training?))

from cascaded-fcn.

@keesh0 , please could you share the source of the new CT volumes you used?
Thank you.

from cascaded-fcn.

Sure.

1. 3DIRCAD dataset-- 20 venous phase enhanced CT volumes from various European hospitals with different CT scanners [test#0, training cfcn network]

2. TCGA-LIHC is The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) [test set]
   https://wiki.cancerimagingarchive.net/display/Public/TCGA-LIHC

from cascaded-fcn.

@keesh0 , very many thanks.
Besides, I encountered some bugs whilst trying out this repo, I have opened an issue on it here:
#34 and I hope the owner will proffer solution to it.
Where you able to plot accuracy curve for the predicted images?

from cascaded-fcn.

I did not try to plot the accuracy curve.
Here is my wrapper code-- https://github.com/keesh0/cfcn_test_inference/blob/master/python/test_cascaded_unet_inference.py

from cascaded-fcn.