We are excited to announce a new public data release from the CMRxRecon2025 Challenge! This update releases the previously unreleased evaluation datasets (https://www.synapse.org/Synapse:syn59814210/wiki/631023) that were originally reserved only for challenge assessment, including the ValidationSet and TestSet, together with corresponding disease labels and radiologist quality control (QC) annotations. 🔹 Released data include: • Full k-space data for the TrainingSet, ValidationSet, and TestSet (TaskR1 & TaskR2) • Radiologist QC scoring sheets • Disease information sheets In addition, the TrainingSet has been re-released to correct a previous issue where the temporal dimension of the flow modality was not placed as the final dimension. This issue has now been resolved. 🚀 The full k-space data, QC annotations, and disease labels for TaskS1 and TaskS2 will be released in a future update. We warmly welcome the community to download and use these data for future research in accelerated cardiovascular MRI, reconstruction foundation models, quality control, and clinically-aware AI. At the same time, our new #CMRx4DFlow Challenge (https://cmrx.chihucloud.com/2026) is also progressing rapidly, and we warmly welcome the community to join us in advancing ultrafast 4D Flow MRI reconstruction and hemodynamic AI research! #MRI #CardiacMRI #MedicalImaging #AI #DeepLearning #CMRxRecon #MICCAI #OpenScience #Radiology #FoundationModels

Created by Recon CMR CMRxRecon
Thanks! That works!
Hi, password can be found at the last line of this file [https://www.synapse.org/Synapse:syn74832951](https://www.synapse.org/Synapse:syn74832951), and you should merge files in order by executing a command like the following, ``` cat `seq -f "tmp-R1/TaskR1-part-%03g.zip" 0 50` > TaskR1-encrypted.zip ```
Hi, I downloaded and merged the zip parts for the updated TrainingData, TaskR1, and TaskR2 datasets from Google Drive. For each dataset, I first combined the split zip files into a single archive (e.g., using `cat ... > *.zip`) and then attempted to extract them with `unzip`. The merged archives seem valid, but during extraction I am prompted for a password for the `.mat` files. I am not sure whether I may have missed an additional extraction/decryption step in the instructions, or whether these archives are intentionally password protected. Could you please confirm whether a password is required for these datasets, and if so, could you provide it? Thank you very much!

📢 New Post-Challenge Data Release for #CMRxRecon2025 page is loading…