/csfm

Code for "Joint Optimization of Hadamard Sensing and Reconstruction in Compressed Sensing Fluorescence Microscopy"

Primary LanguagePython

Compressed Sensing Fluorescence Microscopy (MICCAI 2021)

Code for MICCAI 2021 paper "Joint Optimization of Hadamard Sensing and Reconstruction in Compressed Sensing Fluorescence Microscopy", arXiv.2105.07961.

This repository runs code for the above paper on the publicly-available FMD dataset.

Data

Download FMD Dataset from https://github.com/yinhaoz/denoising-fluorescence and put the root folder in denoising-fluorescent/

Train

Usage

python scripts/run.py -fp <experiment_name> --mask_type <mask_type>

<mask_type> can be one of [learned, random, equispaced, uniform, halfhalf].

Example

python scripts/run.py -fp example --mask_type learned

Model checkpoints and other arguments are saved to out/.

Dependencies

Code was ran on:

  • python 3.7.10
  • pytorch 1.4.0
  • tqdm 4.60.0
  • numpy 1.19.2
  • torchvision 0.5.0