10 July 2020
Our Calgary-Campinas team presented the results of our Multi-channel MR Image Reconstruction Challenge (MC-MRRec) yesterday on the last day of the 2020 Medical Imaging with Deep Learning conference. Although many participants are in very different time zones, over 30 people attended our session, and we have made the video of the session publicly available for people who want to watch it at a later date.
Our challenge had two separate tracks, and it received submissions from 3 teams. Teams that submitted to the challenge received as a token of appreciation early access to longitudinal magnetic resonance imaging (MRI) data that are expected to be made publicly available to the general public in the final months of 2020.
The teams were:
ResoNNance: composed by Matthan Caan, Dimitrios Karkalousos, Nikita Moriakov and Jonas Teuwen. The team members are affiliated with the Netherlands Cancer Institute, the Radboud University Medical Center, and the Amsterdam UMC .
M-L UNICAMP: composed by Lívia Rodrigues, Alexandre Lopes, Hélio Pedrini, and Letícia Rittner. The team members are affiliated with the University of Campinas (UNICAMP).
TUMRI: composed by Shahrooz Faghihroohi, Devendra Vyas, Fabian Gröger, Maik Dannecke, and Viktor Studenyak. The team members are affiliated with the Technical University of Munich.
The winner of both tracks of the challenge was team ResoNNance! Congratulations to the team members for presenting an exciting, innovative, and efficient solution! The team made their code publicly available here!
M-L UNICAMP and TUMRI were ranked second and third, respectively, on the first track of the challenge. Their solutions proposed very interesting concepts that can potentially be used to improve MRI reconstruction.
The challenge goes on as an online challenge, and we will keep receiving submissions and updating the leaderboard.