search

peng-lab / BaSiCPy

MIT License
67 stars 22 forks source link

Input images vs flat-field type #158

Open mbahin opened 4 weeks ago

mbahin commented 4 weeks ago

Hi,

Thanks for your nice tool.

I'm using BaSiC for a few months now and just discovered that the flat-field image is in 32bits. My images are 16bits (and the corrected ones also once processed by BaSiC). Does it have an impact on the quality of the illumination correction? Should the times match?

Cheers, Mathieu

tying84 commented 4 weeks ago

Dear user,

The flatfield is 32 bits as it is a float image that has a mean value of 1.0. but if you use basic build in correction, you corrected image is done by converting your image into 32 bits, divided by flatfield, convert back to 16 bits. You do not need to convert your image to 32 bits yourself

mbahin @.***> schrieb am Mo. 28. Okt. 2024 um 17:08:

Hi,

Thanks for your nice tool.

I'm using BaSiC for a few months now and just discovered that the flat-field image is in 32bits. My images are 16bits (and the corrected ones also once processed by BaSiC). Does it have an impact on the quality of the illumination correction? Should the times match?

Cheers, Mathieu

— Reply to this email directly, view it on GitHub https://github.com/peng-lab/BaSiCPy/issues/158, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACXOKO6VOEAJHH6DGLLXPFTZ5ZHNRAVCNFSM6AAAAABQXZDKCCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGYYTQOBWHA2DMMY . You are receiving this because you are subscribed to this thread.Message ID: @.***>

mbahin commented 4 weeks ago

Hi Tying,

Thanks for the very quick and informative answer!

I use the Python implementation. I guess that since it's a division (I thought it was a multiplication) and not an addition/subtraction, the input format and the flatfield format doesn't have so much impact one on the other anyway.

Cheers, Mathieu

mbahin commented 4 weeks ago

Hi again,

And I have another question. What I was usually doing is to correct the illumination for a set of ~1,500 BrightField tiles (they have a central halo) for one timepoint. This works well. In my datasets, I also have fluo channel and timelapse (both). I’d like to correct the fluo degradation over timelapse. I’ve seen it also in the tutorials but the commands seems to be the same. But I’m not sure whether I should correct each tile individually from all its timepoints (then some immobile stuff disappear, which seems logical) or provide all the dataset but then I don’t know how to format it for BaSiC.

Cheers, Mathieu

tying84 commented 4 weeks ago

That is more complicated. Could you write me an email that we can take a look in zoom?

Best,

Tingying

mbahin @.***> schrieb am Mo. 28. Okt. 2024 um 18:26:

Hi again,

And I have another question. What I was usually doing is to correct the illumination for a set of ~1,500 BrightField tiles (they have a central halo) for one timepoint. This works well. In my datasets, I also have fluo channel and timelapse (both). I’d like to correct the fluo degradation over timelapse. I’ve seen it also in the tutorials but the commands seems to be the same. But I’m not sure whether I should correct each tile individually from all its timepoints (then some immobile stuff disappear, which seems logical) or provide all the dataset but then I don’t know how to format it for BaSiC.

Cheers, Mathieu

— Reply to this email directly, view it on GitHub https://github.com/peng-lab/BaSiCPy/issues/158#issuecomment-2442198922, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACXOKO6TJ7PTP3G7YO3CYUTZ5ZQVBAVCNFSM6AAAAABQXZDKCCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINBSGE4TQOJSGI . You are receiving this because you commented.Message ID: @.***>

mbahin commented 4 weeks ago

Hi Tingying,

I don't know where to find your email address. I added mine to my profile, feel to contact me so that can arrange.

Cheers, Mathieu