Unable to reproduce results in register_ihc.py example

[changwxx]

Hi! Thank you for your awesome work and for sharing the code. I've encountered an issue where I can't reproduce the same results as those stored in examples/expected_results/registration/ihc. I followed the instructions and ran examples/register_ihc.py without making any changes. Could you please let me know if I missed something?

I found the masks are not good enough.

And the ihc_non_rigid_overlap and ihc_rigid_overlap registration results I'm getting are different from the expected ones shown below.

[cdgatenbee]

Hi @changwxx,
Thanks for reporting this. It does look like mask for 2nd image isn't quite right. I'll take another look at the examples and see if I can reproduce this.

Best,
-Chandler

[cdgatenbee]

Hi @changwxx,
I re-ran the examples on MacOS using Python 3.9-3.11, but haven't been able to recreate this issue. Could you try using the Docker container [here] to see if it works correctly for you?

Best,
-Chandler

[aschiebub]

Hello,

I have the exact same issue using the Docker container to run register_ihc.py (the masks produced are exactly the same as changwxx). Moreover, I've been trying to use Valis on my own dataset, and the masks are also incorrect on several samples, leaving behind fair amounts of tissue.

I would highly appreciate your help on this matter!
Thanks

[changwxx]

Hi @cdgatenbee!

Thank you for your response. I tried using the Docker container and re-ran the examples/register_ihc.py, but I am still encountering the same masking issue as before. Could you please help me address the issue? Are there any specific parameters that need to be passed into the register() function, as seen in this line of code?

Thanks for your help!

[cdgatenbee]

Hi @changwxx and @aschiebub,
Apologies for the delayed response. I was able to recreate this issue using the Docker container, so I can use that to debug this masking issue. I'll be working on it and will let you as soon as I've figured out the solution.

Best,
-Chandler