CBICA / CaPTk

Cancer Imaging Phenomics Toolkit (CaPTk) is a software platform to perform image analysis and predictive modeling tasks. Documentation: https://cbica.github.io/CaPTk
https://www.cbica.upenn.edu/captk
Other
180 stars 64 forks source link

Brain Tumor Segmentation #529

Closed jacobantonyalappatt closed 4 years ago

jacobantonyalappatt commented 5 years ago

Describe the bug After providing the necessary image contrasts (T1, T1ce, FLAIR, T2), it throws this error message "DeepMedic returned with exit code !=0". Previously, I have used the DeepMedic executable to segment the tumor in this particular patient, using these 4 co-registered images. As these images had worked before, and I have the segmented tumor image from DeepMedic, I wanted to compare CaPTk's output, to see if there are differences being introduced.

To Reproduce Steps to reproduce the behavior:

  1. I provided 4 co-registered images.
  2. They succesfully overlaid in the viewer, and I selected the appropriate modality for each of them.
  3. I selected Deep Learning > Brain Tumor Segmentation
  4. Allowed the machine to select the existing model directory, and I specified a new output directory.

Behavior It is surprising that an error was thrown by CaPTk after I provided the same 4 images. Error report also fails to point the user in a direction that would help them solve the issue, if it is with respect to the images.

CaPTk Version The installed version is 1.7.0, but in the Help>About section, all I can see is a blank black background. There is no visible content in the 'About CaPTk' popup box.

Desktop:

sarthakpati commented 5 years ago

I believe this issue has been resolved in the builds that originated in the previous week. You can try the latest here (accessible only via the UPHS network): S:/CBICA/CaPTk/installers/CaPTk_1.7.0_nonRelease_Installer.exe

Closing the issue for now. Please reopen if it persists.

dboun commented 4 years ago

Are we sure this is resolved? Because last time I tried the Deep Learning Brain Tumor Segmentation it didn't work, and also someone at RSNA mentioned the same thing

sarthakpati commented 4 years ago

I am not able to replicate this on 1.7.2 or any of the development builds. I feel that the person who mentioned this at RSNA has probably something funky going on with the data. Will need more context regarding the data being used. Once I get that, I'll open up a new issue.