Submission

Submissions will be accepted from 5 Dec. 2023 and will continue to accept until 15 Dec. 2023. Please make sure that your code adheres to the attached format to ensure easy reproducibility of the results on our end. In addition, we also request you to send us a link to your GitHub repositories (if private, kindly add us as collaborators. AnderBiguri (ander.biguri@gmail.com) and Subhadip-1) with clear instructions to run your code, together with a conda environment file so that we can install the required dependencies. You are also requested to share with us a two-page summary of your method, explaining the network architectures, training details, etc.

Please send the above materials over email to us by no later than 15 Dec. 2023.  

We will evaluate the submissions after the 15th of Dec. and the result will be announced shortly thereafter. We will keep you posted about the next steps.  

We look forward to receiving your codes.  

Python template script:

#%% Challenge imports

import torch

import numpy as np

##################################

#%% Your imports

##################################

#%% Load data

folder="test/folder/in/organizers/path"

# Provide a Dataloader functio or loop to iterate over files in folder

# e.g.

dataloader = my_data_loader(folder)

# You can assume that the data format, filenames and folder structure of the test will be the same as the training data.

#%% Main loop

# Either pytorch loop, or your own

# Option A:

for sinogram, target_reconstruction in dataloader:

    reconstruction= my_recon(sinogram)

# Option B:

for noisy_reconstruction, target_reconstruction in dataloader:

    reconstruction= my_recon(noisy_reconstruction)

# Option C:

# Provide your own loading loop that is not based in pytorch dataloader.

##################################

#%% Notes:

# 1- Provide a Conda package. We can test it in Linux and Windows but we have limited GPU resources in Linux. Please specify the OS tested on.

# 2- Provide a README.md file with instructions on how to run your code.

# 3- Be available by email in the folliwing weeks. If we can't run your code we will contact you.

# 4- If you have any questions please contact us.

##################################