Prep by Lacy Barton.
Zeiss Light Sheet Z1 image processing notes
Macros for processing live embryo lightsheet datasets is LightsheetZ1_macrosMC_v109_nomaxp.txt
Please keep in mind this is our first shot at both the live microscopy & the image processing.
At each step, important to manually check that all files were really processed. If 303 original files, scan down the directory listing to check all 303 were processed.
1.
"Open images, process, and save results [q]"
The first step was to rescale the XY size in a series of Z series taken through time. In our case, there were two channels and between 300 and 400 Z slices at each timepoint (every 2 or 4 minutes for 10 to 13 hours). This made each stack 1/4 the original size and reduced the camera noise (we may try flatfield correction in futre datasets).
Problem: LOCI plugins sometimes reads first image in sequence as not a single files but as the entire dataset.
Solution: Add character to filename so is does not match the rest of the sequence. After processing, manually rename the first file and the resultant file(s).
Also, a maximum projection was generated and saved, but this should be moved to the next step after volume subtraction.
2.
"sub one channel from another, series off disk"
The next step was to open each of the resultant TIF stacks, split the channels, and subtract the green (channel 1) from the red (channel 2) and save only the result. Before doing this:
> need to put all the files in a dedicated folder and
> need to check method manually with at least one representative timepoint to see if other math necessary such as subtracting 50 from the green and then multiplying by 2 to adjust the background and the image before subtraction.
3.
"open crop project save"
Based on empical playing by the user, a box is set in XY for cropping the volume and projection settings are set. Other things could be done too.
The input folder should only have the subtracted stacks from step 2 or other processed stacks.
This movie is at 25% spatial scale in XY (12.5% of the raw data) of the first 78 timepoints of 210 timepoints made using this method.
Prep by Lacy Barton.
These steps could be combined in a single pass, but we have to define a standard series of operations before steps can be combined.
comments, questions, suggestions: Michael.Cammer@med.nyu.edu