Leader Board

General Information

The aim of the Super-resolution of Multi-dimensional Diffusion MRI (Super MUDI) Challenge is to super-resolve combined diffusion-relaxometry MRI data.

The challenge consists of two tasks. Each task explores a different low resolution MRI acquisition strategy, leading to two different types of images to super-resolve. Only submissions that attempt both of the tasks will be considered for evaluation.

The aim in Task 1 is to super-resolve data that has high in-plane resolution but thick (axial) slices, whereas Task 2 consists of super-resolving data with isotropically lower resolution. The images below show an example of the downsampling procedure in each task.

The winner of the challenge will be the submission with the overall best score (lower value) across the two tasks, computed as the sum of the scores in each task (see individual task description for further details about the metrics used for evaluation).

In this way, we will be able to draw conclusions on two aspects of super-resolution: 1) how reliable and stable is a super-resolution method; 2) which combination of sub-sampling strategy and super-resolution method is the best alternative to apply in a clinical scenario. The richness of contrasts in the MUDI data will allow us to generalize the derived conclusions to many MR imaging modalities.

The challenge results and discussion will be presented within the CDMRI'20 full-day workshop.

PLEASE NOTE! To evaluate your results, we will only be providing subsampled diffusion data.

Get the Data

Data is provided by the Centre for the Developing Brain, Centre for Medical Engineering, King's College London, UK. All interested participants need to fill out the data sharing agreement (OA-Agreement) and the interest form that can be found here and the data will then be sent via file link.

Please send the filled OA-Agreement form and any questions you may have about the challenge to cdmri20@cs.ucl.ac.uk

Super-MUDI Challenge 2020 - Instructions

By filling out the data sharing agreement, you will be able to download at least 5 different test subjects for task 1 and for task 2: the test subjects will be the anisotropic (task 1) and isotropic (task 2) low-resolution version of datasets of which we have the groundtruth (not accessible to participants). You will need to super-resolve them to the original isotropic 2.5x2.5x2.5mm resolution.

The data was acquired with the same protocol as in the MUDI Challenge from CDMRI'19

The deadline for submitting your results is September the 15th, 2020.

Super-MUDI Challenge 2020 - Evaluation

We will evaluate the submissions using the Mean Squared Error (MSE) between your (submitted) super-resolved datasets and our ground-truth. The lower the sum of the MSE over the 2 tasks (whole-brain mask and averaged across the test subjects) the better your ranking. Easy! We will however analyze regional differences and evaluate the impact of the different MRI acquisition parameters on the overall set of submissions from participants.

PLEASE NOTE! To evaluate your results, we will only be providing subsampled diffusion data.

Important dates

28th of May, 2020 Data release
September 1st, 2020 Indicate intention to submit
September 21th, 2020 Challenge submission deadline
October 2020 Challenge and workshop in Virtual