Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Bug] TotalSegmentator dataset taking longer to load compared to v2 #3523

Closed
igoroctaviano opened this issue Jul 6, 2023 · 15 comments
Closed
Labels
Awaiting Reproduction Can we reproduce the reported bug? IDC:candidate Possible feature requests for discussion, and if we agree, they can be relabeled IDC:priority

Comments

@igoroctaviano
Copy link
Contributor

Describe the Bug

Based on recent tests with v2 and v3 we identified that the loading of the segmentations are faster in v2.
This issue should further investigate and address the problem.

  • TotalSegmentator dataset shared with @sedghi

Steps to Reproduce

  1. Load TotalSegmentator dataset
  2. Load segmentation series (benchmark v2 and v3)

The current behavior

Loads slower in v3

The expected behavior

Loads better or equal to v2

OS

macOS 13.4 (22F66)

Node version

16

Browser

Version 114.0.5735.198 (Official Build) (x86_64)

@igoroctaviano igoroctaviano added the Awaiting Reproduction Can we reproduce the reported bug? label Jul 6, 2023
@igoroctaviano igoroctaviano added the IDC:candidate Possible feature requests for discussion, and if we agree, they can be relabeled IDC:priority label Jul 6, 2023
@sedghi
Copy link
Member

sedghi commented Jul 10, 2023

I can't run the total segmentor case in OHIF v2, which study you are trying to load?

image

@igoroctaviano
Copy link
Contributor Author

@sedghi
Copy link
Member

sedghi commented Jul 10, 2023

Were you supposed to share a screenshot or a dataset?

@igoroctaviano
Copy link
Contributor Author

Were you supposed to share a screenshot or a dataset?

Remember the drive I shared with a bunch of datasets e.g. the RGB one? there you can find the totalsegmentator dataset.

@sedghi
Copy link
Member

sedghi commented Jul 11, 2023

This file doesn't have a SEG

image

I'm using another study which you can find here, but I'm not able to load it in v2 either

https://www.dropbox.com/t/eEseMevPSwqJLsDt

@igoroctaviano
Copy link
Contributor Author

@sedghi I just uploaded another zip with the seg file to that drive.

@igoroctaviano
Copy link
Contributor Author

igoroctaviano commented Aug 17, 2023

@sedghi #3577 fix was tested using zip I uploaded or other dataset?

@rodrigobasilio2022
Copy link
Collaborator

rodrigobasilio2022 commented Aug 23, 2023

It seems to be fixed as stated in the title: old v3 version (~33s) new v3 version (~11s). I couldnt open the cited study in v2 as it gives the following error

image

@rodrigobasilio2022
Copy link
Collaborator

rodrigobasilio2022 commented Aug 23, 2023

@fedorov do you have another TotalSegmentator studies to test, to try to understand if this is an isolated problem?

@rodrigobasilio2022
Copy link
Collaborator

rodrigobasilio2022 commented Aug 23, 2023

Using the link: https://viewer.imaging.datacommons.cancer.gov/viewer/61.7.186150680759207515465616855509993541745
The study opens in v2, but the comparison is bogus as the segmentation in IDC only has 13 segments while the segmentation in google drive has 98 segments. Also in v2, despite the segmentation appears to load fast the user cannot manipulate the image as it is loading the main series images

Using exactly the same images for both viewers, the loadign appears to be very similar

@rodrigobasilio2022
Copy link
Collaborator

@fedorov, do you want to investigate why v2 didnt load the full segmentation?

@fedorov
Copy link
Member

fedorov commented Aug 23, 2023

The study in IDC contains only the segmentation that was loaded. It does not have the same segmentation with 98 segments shared separately.

@rodrigobasilio2022
Copy link
Collaborator

The file shared with Ali Reza has it. Anyway, problem seems to be solved.

@fedorov
Copy link
Member

fedorov commented Aug 25, 2023

To confirm, tests O16 and O19 in the IDC test inventory (access restricted to IDC project participants) should be used to test this functionality.

@fedorov
Copy link
Member

fedorov commented Aug 25, 2023

I confirmed performance is very good now, nice work @rodrigobasilio2022! 👍

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Awaiting Reproduction Can we reproduce the reported bug? IDC:candidate Possible feature requests for discussion, and if we agree, they can be relabeled IDC:priority
Projects
Status: Done
Development

No branches or pull requests

4 participants