SRRF-CUPY

An implementation of super-resolution radial fluctuations(SRRF) with CUDA support.

Overview

This program helps improve the resolution of fluorescent microscoping, based on the statistic properties of fluorescent fluctuation.

Here is the paper

Installation

edit cupy-cuda100 in requirements.txt to fit your CUDA version.

for CUDA 8.0, change it to cupy-cuda80
for CUDA 9.1, change it to cupy-cuda91
for CUDA 10.1, change it to cupy-cuda101

You can reference installation guide of CuPy.

Then

pip install .

Usage

GUI

After you have all dependencies installed, run

python ./srrf_cupy/ui_srrf.py

or

python -m srrf_cupy

if GUI fails to start, a simple CLI version for debug will start instead.

Import as a module

import srrf_cupy.srrf as srrf

# read data from a folder
img_array = srrf.read_images_from_folder(folder_path)     

# or read data from a tiff file
img_array = srrf.read_images_from_tiff_file(file_path)    

# run SRRF with default args
ret_img_array = srrf.srrf(img_array)

# show result
srrf.plot_result(img_array, ret_img_array)

Reference:

[1] Han, Y., Lu, X., Zhang, Z., Liu, W., Chen, Y., Liu, X., Hao, X., and Kuang, C. (2019). Ultra-fast, universal super-resolution radial fluctuations (SRRF) algorithm for live-cell super-resolution microscopy. Opt Express 27, 38337-38348.