search

RobertMolenaar-UT / fastFLIM

PicoQuant PTU file to fastFLIM images
GNU General Public License v3.0
1 stars 0 forks source link

Bidirectional scanning PTU sample files? #1

Closed cgohlke closed 2 days ago

cgohlke commented 5 days ago

Hello. I noticed that your package supports PTU files acquired with bidirectional scanning. I am trying to support decoding bidirectional scanning in the ptufile Python library, but have been unable to obtain any such files for testing. Would you be able to share such files? Thank you.

RobertMolenaar-UT commented 5 days ago

Hi Christoph , I try to find you a nice picture measured with the Bidirectional scanning this week. Note that when you use Flimbee scanner in bidirectional mode, there is some hysteresis between the odd and even lines (that is scan rate dependent) in experiment you need to set it correctly. However that is not always correct. So in ReadPTU-bidirect Is pass a bi-shift variable. That works at pixel value, but perhaps I need to change it to subpixel.

I will try to find you this week one or some examples. FYI information I modified PTU_reader 3 months ago, not all PTU header data was passed, now it does. Known feature bug, it cannot extract single frames measured from a multi-frame image. useful for kinetics. Robert

cgohlke commented 2 days ago

Thank you for privately sharing the sample files!

cgohlke commented 2 days ago

Hi Robert,

I am now able to decode bidirectional-scan PTU files and can reproduce your findings on hysteresis between the odd and even lines. Two of the files you shared require about half a pixel shift correction. The third file looks good visually without correction. The best shift can probably be found by correlating images from even and odd lines.

There is a related entry in the FLIMbee manual: """ Bi-directional scanning: The software tries to correct the offset between the lines scanned from left to right and the lines scanned from right to left. However, in most cases, there is a slight mismatch of a few pixels (most visible if horizontal stripes are present in your image). When clicking on the “More” button, the bi-directional offset can be adjusted to get rid of this mismatch. At very high scan speeds, the offset might be different in selected regions in the image. This artifact cannot be corrected for. """

I don't understand why the shift correction is left to the users in the first place. Can't the line markers be correctly set in the PTU stream?