AUTHOR: (Karl) Sangwon Lee
UPDATED: May 27, 2021
Scritps Written for BIOCARD Project at Miller/Tward Lab
This is a czi-tif conversion script that reads a .czi image, stack its RGB channel, and export it as tiff file.
Previously, BIOFORMAT importers were unable to read .czi image in headless console mode due to its inherent incapability.
This script bypasses it using a unique CZI reader library within the script.
Resulting .tif file is about half the file size of .czi file, while keeping the resolution the same.
Thanks to sebi06 from Carl-Zeiss Microscopy for the help with this script.
This is a bash shell script written to automate applying of czi conversion to all .czi files within a directory.
The thing is, ImageJ is not so nice in that headless mode script seems unable to take more than one input parameter.
Meaning, the user could not specify input directory AND output directory.
This script is written to bypass that by directly editing the script using bash commands. Quite primitive - but it works.
Repeat the code block indicated in the bash script and edit the path to apply conversion to multiple folders.
This is a bash shell script written to automate applying of ZhenZhen's UNet apply.py to all .tif files within a directory.
I suggest using the following shell command to use this:
nohup ./unet_script_brain4.sh &
This will run the shell script in background.
It seems that you cannot run pytorch sessions in parallel (maybe there's a way?), so what I am doing here is running one unet after another via this script.
Similar to the shell script above, repeat the code block indicated to apply unet to multiple folders.
brain4 is just an arbitrary naming. I copied the same script and made it brain5, brain6, etc so thath I would not have to edit same file over for different brains.
This is a bash shell script written to automate applying of ZhenZhen's feature extraction to all .npy probability map files within a directory.
This will submit all feature extractions for each file together, and they will run in parallel.
Feature extraction takes a long time, depending on the number of tau particles and number of files. It could take as long as 7 days or more if lots of files.
To monitor your script process in shell, use the following:
ps -aux
This will show all the running processes, and you should be able to find rows of feature extraction running for your user node.
Feel free to email me (slee508@jhu.edu) for any questions regarding the scripts