Hello everyone,
I am trying to use the new Micrograph Denoiser and see these circles in the middle of my micrograph (with default setting). I am puzzled what are these, they are not present in my original micrograph after patch motion correction. I was wondering if I can change some parameters in input settings to fix these.
Anamika
Were the micrographs you posted produced with Micrograph Denoiser default settings?
Yes they were produced with default settings.
@Anamika1 If you push the Show from top button under Event Log, you should be able to scroll down to a section with learning progress plot and two (split) panels of micrographs like
Please can you post a screenshot of that for your example.
I think in my case the training set does not match the validation set. Any suggestions to fix this issue.
Please can you provide details that make you think so?
I was just saying that based on comparing the plot you posted vs what my data plot look like in the post above.
Hi @Anamika1!
Did your collection facility provide you with a defect file? If so, did you provide it during motion correction?
Also, I wonder if you could share with us an example of your motion correction results (like those below):
![image](https://discuss.cryosparc.com/uploads/default/original/2X/6/60b09510754c2dafadb451adfe333bc5acd3dd82.png)
Thank you!
Hi, Please see below the example of motion correction file for this data and also I did not use the Detector defect file. By the way this data was collected on falconIV detector.
Hi @Anamika1! Thanks for that info!
Do all of the diagnostic plots have this region of patches with little or no movement?
![image](https://discuss.cryosparc.com/uploads/default/original/2X/3/3181e729d2006a5d73eb679c8e963224bf48a2fe.jpeg)
Hi @Anamika1!
I believe this problem results from a hot pixel. When we apply the CTF to the micrograph before denoising it, a hot pixel will essentially recapitulate the CTF, resulting in the rings you can see there. I believe the easiest way to fix this is to reach out to your collection facility with this information and ask them for a defect file.
This hot pixel also seems to be affecting your motion correction results (note that the stationary patches are in the same place as the rings). I would therefore recommend getting a defect file even if you decide not to move forward with denoising.
I wish I had an easier fix for you, but I hope that is helpful!
Thanks, this make sense. I will try to process with defect file and see if that resolve the issue.
Anamika