search

ENHANCE-PET / MOOSE

MOOSE (Multi-organ objective segmentation) a data-centric AI solution that generates multilabel organ segmentations to facilitate systemic TB whole-person research.The pipeline is based on nn-UNet and has the capability to segment 120 unique tissue classes from a whole-body 18F-FDG PET/CT image.
https://enhance.pet
GNU General Public License v3.0
196 stars 28 forks source link

Unable to download clin_pt_fdg_tumor with both Moose v2.0, v3.0 #149

Closed HFahmida closed 1 month ago

HFahmida commented 1 month ago

Hi, I wanted to use the moose FDG-PET tumor segmentation model. However, I see it is not present in V3.0. I installed V2.0, and I was not able to download the model weight as the website link in the resource.py file, seems not a real website.

image

resource.py website link for V2.0:

Screenshot 2024-10-09 113402

I also tried V2.2.30, and it says "File is not a zip file". The Error:

error_moose2

resource.py file with the website link to download the model weights:

image

I am looking for help to have the model weights for the "clin_pt_fdg_tumor" model. Thank you.

LalithShiyam commented 1 month ago

Hi there, we removed the tumor model, because it's an altogether a new tool. Sorry :) it won't be in moose anymore.

HFahmida commented 1 month ago

Hi, Thank you for promptly responding to my queries. Will this model be publicly available?

LalithShiyam commented 1 month ago

Absolutely; we will release the tool soon. But if you sign up for enhance, you can get it even earlier during beta testing.

Https://enhance.pet

harmonsa commented 1 month ago

Hi, does this mean we cannot use the tumor model from older versions (2.0) of moose as well? The link to the weights does not exist therefore the older versions of moose now do not work. We would like to validate our results against the moose (2.0) FDG tumor model.

LalithShiyam commented 1 month ago

Hi there,

Unfortunately not. We are publishing our own paper as well and it's not fully trained yet. And since we haven't published our previous model, it wouldn't make a lot of sense to benchmark against us...

And the previous model was just there for a small task therefore, it hasn't been published in a real journal.

Cheers, Lalith

LalithShiyam commented 1 month ago

Closing this due to inactivity.