Enhancing high-content imaging for studying microtubule networks at large-scale

Deep neural network to turn noisy widefield images into clean confocal images

published in Machine Learning for Healthcare 2019

1. simulations: Scripts and simulated data for performance benchmarking
2. demo: Scripts of how to use CycleGAN models
3. checkpoints: Some pretrained models

Requirement:

Scripts are developed using Python 3. CycleGAN will need to be set up properly. See comments in the notebook.