search

abria / TeraStitcher

A tool for fast automatic 3D-stitching of teravoxel-sized microscopy images
http://abria.github.io/TeraStitcher/
Other
82 stars 32 forks source link

Missing slices in merging #69

Open sharonkatz510 opened 4 years ago

sharonkatz510 commented 4 years ago

Hey, I tried to stitch a dataset and got 2 channels great. The 3rd channel however, seemed to work fine during all stages (the progress bar said 783 slices were merging) but resulted in a stack with only 588 slices. This problem repeats only with this specific channel. The "merging.xml" file says it should output a 783 slices stack (based on the "ABS_D" and "Z range" parameters). Also, i made sure and all the stitched stacks have the same number of slices (800). Could you think of an idea for further debugging? Has this kind of bug ever occurred to you?

Added the xml_merging.xml as .txt for inspection. xml_merging.txt

Thanks!

iannellog commented 4 years ago

I need more information to help you. I understand you have the channels stored in separate files (is this correct?) and perhaps you have stitched them separately, including the alignment computation step. If this is the case, it not the right way to proceed. Please have a look to section 1.8 of the attached guide and then try to send me as much information as possible about which is your case among those described in the guide.

TeraTools documentation (draft 200204).pdf

sharonkatz510 commented 4 years ago

Hi Giulio, thanks for the quick response! I have 3 homologous channels that are stored in separate directories . My desired result are 3 stitched Tiff stacks- one for each channel. I'm running TeraStitcher with the CLI version, and an xml_import file that i generate for the best detailed channel according to the TeraSrtitcher format. According to the documentation, I first choose the best detailed channel, run the entire stitching process and examine the stitched stack. If the stitching is good, I then manually change the xml_import and xml_merging files to suit the 2nd channel directory and file names. I copy those files to the 2nd channel directory and run the import step (to receive a new metadata file) then i run the merging command (I do not run the computation steps). I repeat this process for the 3rd channel as well. This method usually works fine but for some reason in the current dataset I get a different number of Z slices in the 3rd channel (588 instead of 783 in the other channels). It is worth mentioning that during merging of the 3rd channel the progress-bar prompted it was merging 788 slices (as expected but not received).

I will be happy to give any other information that might be relevant, just say so.