'A Bit Too Much? High Speed Imaging from Sparse Photon Counts' Python Implementation

This repository is a python implementation of the scheme proposed in

Chandramouli, Paramanand, et al. "A bit too much? high speed imaging from sparse photon counts." 
2019 IEEE International Conference on Computational Photography (ICCP). IEEE, 2019.

A preprint can be found at https://arxiv.org/abs/1811.02396.

Code

The code is structered into data loading (dataloader.py, torch_augment.py, torch_augment_functions.py), definition of the network architecture (network.py), specification of the loss function (loss.py) and auxiliary files to define the training and testing procedure (solver.py, scheduler.py) as well as code for n-dimensional stitching (torch_stitching.py). Some exemplary network snapshots can be found in /snapshots.

Requirements

• PyTorch
• Scikit-image
• H5py

The exact setup can be installed by running

conda env create -f environment.yml
conda activate HighSpeedImaging

Data

The dataset used for training and testing of the network is the Deep Video Deblurring for Hand-held Cameras Dataset which is publicly available.

This code has been developed together with mj9 as part of a university project.