Join a competition to train and refine AI models for cancer research
Published on 14/12/2023
The EU-funded project CHAIMELEON, which is a member of the Artificial Intelligence For Health Imaging (AI4HI) cluster as INCISIVE, has launched the ‘CHAIMELEON Open Challenges’, a unique opportunity for European scientists to advance cancer research with AI, in compliance with the strictest data privacy regulations and ethical standards. The project encourages participants to collaborate in developing and training new AI-powered solutions to improve the diagnosis and treatment of prostate, lung, breast, colon, and rectal cancers.
As the project explains on its website, by joining the CHAIMELEON Open Challenges, researchers will have access to quality datasets, including radiological medical images and clinical variables, to train their AI models.
Competition phases:
The challenges will be divided into two phases: the Classification Phase and the Championship Phase.
The Classification Phase is open until December 31st, 2023. In this phase, any participant can join and download a highly controlled cancer dataset. The dataset will consist of post-processed (i.e., harmonised) images associated with specific clinical variables, ensuring a level playing field for all participants. The participants will train their AI models and compete to be among the top on the leaderboard.
The Championship Phase will extend from January 1st 2024 to February 29th, 2024. In this phase, the top 40 participants from the previous phase will be invited to compete on an even grander stage, using the CHAIMELEON platform to train their algorithms. This phase will include five challenges, one for each cancer type (prostate, lung, breast, colon, and rectal), with a total prize pool of 20.000€ per challenge, split among the winners. Original images and associated clinical data, along with different image pre-processing and harmonization tools, will be available to participants who have passed the qualifying round cut-off point. The same computational resources will be assigned to each participant; therefore, each participant can decide which challenges to tackle and apply their strategies accordingly.
