FLUMI is a straightforward fluorescence microscopy simulator that uses point-clouds as ground truth structures and returns a simulated z-stack either as a TIFF stack or as PNG files in a directory. The simulation is done in five steps:
- Fluorophores from the point-cloud are seeded onto a voxel grid
- Experimental PSFs are approximated as a Gaussian to allow separable (and memory-efficient) convolution
- Fluorophore photon emission is assigned by drawing from a Poisson distribution
- The voxel grid is convolved with the approximated PSF by compartmentalising the grid into smaller boxes where the convolution is sequentially computed thus returning a simulated sample.
- The simulated sample is ‘acquired’ by binning the voxel grid to match the detector resolution and noise is added:
- The sensor detection accuracy is modelled as a Poisson distribution
- In the analog-to-digial-conversion step a user-defined gain is applied
- Additive electronic noise is modelled as a Gaussian where the DC-offset and sigma can be set by the user
FLUMI can be executed from the command line allowing the user to specify the following arguments:
- input structure (string):
--struc(required) - target location (string):
--save_as(required) - convolution method (either 'box' or 'additive'):
--method - experimental PSF in TIFF format (string):
--psf - numerical aperture (float):
--NA - wavelength (int):
--wl - gain (int):
--gain - DC-offset (int):
--dc_off - sigma noise (float):
--sigma - voxel grid size which is the pixelsize of the input structure (int tuple of shape (x, y, z)):
--voxel_size - PSF size which is also the pixelsize of the simulated sensor (int tuple of shape (x, y, z)):
--psf_size
A simulation using a point-cloud of fluorophores and an experimental PSF that is saved to a directory would look like this:
python main.py --struc 'segmentation.txt' --psf 'PSF_BW.tif' --save_as 'Simulation'
FLUMI runs on Python 3.7.4 and backward compatibility cannot be guaranteed.
pip install tifffile
pip install imageio
pip install argparse
pip install tqdm
pip install numpy
pip install tensorly
pip install scipy
Otherwise create a new conda environment:
conda env create -f flumi.yml
git clone https://github.com/krentzd/flumi.git
Navigate to the folder called flumi and run:
python main.py --struc 'segmentation.txt' --psf 'PSF_BW.tif' --save_as 'Simulation'