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

2d auto segmentation #384

Open
Nal44 opened this issue Feb 7, 2024 · 5 comments
Open

2d auto segmentation #384

Nal44 opened this issue Feb 7, 2024 · 5 comments

Comments

@Nal44
Copy link

Nal44 commented Feb 7, 2024

Hi,

I am testing 2d .

the auto segmentation runs, but doesn't give any outputs (with default parameters) ... ?

Any ideas ?

thanks
antho

@Nal44
Copy link
Author

Nal44 commented Feb 7, 2024

I can share an image if that helps ?

@constantinpape
Copy link
Contributor

If you're using the vit_b_lm then you need to set lower values for the thresholds:

pred_iou_thresh = 0.65, stability_score_thresh = 0.75

We are working on setting these thresholds automatically, see #337.

@Nal44
Copy link
Author

Nal44 commented Feb 8, 2024

Thanks a lot,
using vit_b_lm and pred_iou_thresh = 0.65, stability_score_thresh = 0.75 as suggested :

see below :
I think it needs it might need pre filtering or some normalization step ?
automatic

Any ideas ?
thanks

@constantinpape
Copy link
Contributor

Yes, I think the cells are too small for the current settings of the automated segmentation. I think it is more promising if you follow the approach outlined in #383. For this the automatic segmentation results don't really matter. As long as manual annotation either via point or bounding box prompts works well the refinement approach should work.

@Nal44
Copy link
Author

Nal44 commented Feb 8, 2024

ok, no probs,
working on it :)

(For info ,this is a standard 20x objective image that we use for staining in the lab)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants