jnkather / msifromhe Goto Github PK

The color normalization toolbox does not seem to be accessible without a Warwick account. Unfortunately, I cannot create an accout by myself. Is there a way this is included here in the code?

I am trying to reproduce the results and I found following function "moveImagesToFolder" in step_05_split_train_test.m is not provided .

moveImagesToFolder(cliniPatient,msitrain,cnst.tileInputPath,cnst.tileOutputPath,'msitrain');
moveImagesToFolder(cliniPatient,muttrain,cnst.tileInputPath,cnst.tileOutputPath,'muttrain');
moveImagesToFolder(cliniPatient,msstrain,cnst.tileInputPath,cnst.tileOutputPath,'msstrain');

moveImagesToFolder(cliniPatient,msitest,cnst.tileInputPath,cnst.tileOutputPath,'msitest');
moveImagesToFolder(cliniPatient,muttest,cnst.tileInputPath,cnst.tileOutputPath,'muttest');
moveImagesToFolder(cliniPatient,msstest,cnst.tileInputPath,cnst.tileOutputPath,'msstest');

Where can I find it? Thank you.

I tried to reproduce your work , but I couldn't find the function getTrainingOptions in the step_01, not mentioned in the tool boxes and also the matlab documentation.

I am new to the matlab deep learning area.
Could you tell me how to find the function or how to define it?

run log (Matlab r2018a update 6):

Warning: Directory already exists. 
> In step_01_train_tumor_normal_classifier_final_crossval (line 22) 
*******
STARTING CROSSVAL 1
sucessfully loaded&modified network
Undefined function or variable 'getTrainingOptions'.

Error in step_01_train_tumor_normal_classifier_final_crossval (line 76)
opts = getTrainingOptions(hyperparam,[]);

I am trying to reproduce the results and I found following function is not provided in the subroutine.

currMPP = Info2MPPrescue(currImInfo,ni(end),... % extract MPP; if fail try to rescue
cnst.rescueFolder,strrep(strrep(currImName,'.svs',''),'.tiff',''));

Though I can use Info2MPP() which I found in the subroutine, but I think it does not help on many WSIs from the TCGA-COAD (colon).

Hi kanther, could you let me know more about clinical data as there are may attributes that need more clarity.

Hi, I have found 'splitEachLabel.m' and 'augmentedImageDatastore.m' are missing in 'subroutines'.

I am trying to reproduce your work , but predIsMSIMUTScore is not defined in

On the website provided in the paper, https://zenodo.org/record/2530835#.Xxh1KCgzZHY, I noticed that all image tiles are either labeled as MSI or MSS. I am wondering if it is the case that, if a patient is marked as MSI in TCGA, then all his/her image tiles are labeled as MSI, and vice versa. Is it possible that a patient can have some image tiles that are labeled as MSI while others are labeled as MSS?