search

wildrootlab / SHARCQ

Tools to register histological section images and cell count locations to mouse brain atlas
24 stars 9 forks source link

Issues with ROI Warping #11

Open marhmue opened 1 year ago

marhmue commented 1 year ago

I have not yet been able to make it through the entire SHARCQ paradigm with my coronal images. During the Image-ROI-Atlas-Composite stage (I've tried with both the Allen and Chon atlas), recognised cells seem to be misaligned. I've spent a few days troubleshooting, to no avail. I've looked through raw and transformed cell count/coordinate data, ROI files, transformed ROI files etc. both for my image sets and in comparing these to your test images provided. I've attached a screenshot which shows the error I'm encountering.

Is this something which you've seen or had happen in the past? Any advise or insight as to how this might be fixed would be deeply appreciated @krislauridsen @anni3ly. Thank you for your time and consideration.

Registration_Error_MM_SHARCQ_Inquiry

Originally posted by @marhmue in https://github.com/wildrootlab/SHARCQ/issues/6#issuecomment-1449133546

krislauridsen commented 1 year ago

Hi, have you been able to run the program with the same images all the way through?

marhmue commented 1 year ago

Thank you @krislauridsen - yes, I have been able to run the rest of the program. All other steps seem to be functioning well. I was also able to run the entire pipeline using the test images provided in your publication - it is only when I try my images that I seem to be encountering these issues with (I'm guessing) scaling/warping.

anni3ly commented 1 year ago

Hi @marhmue, it looks to me like the medial-lateral tilt, and perhaps the dorsal-ventral tilt, in the ROI coordinates is off. When you plot the X,Y coordinates in the csv file of ROI coordinates, does it look like the green dots in the screenshot you shared?