CMR×Recon

CMR×Recon数据集是由复旦大学人类表型组研究院等机构联合创建，包含来自300名健康志愿者的多对比、多视角、多切片和多线圈心脏磁共振成像数据。该数据集涵盖心脏电影和映射序列，并提供了所有受试者心肌和心室的精细分割。数据集旨在通过提供标准化评估标准和公开访问，推动心脏磁共振成像重建技术的发展。创建过程中，使用3T扫描仪和32通道心脏线圈进行数据采集，并通过专业软件进行数据处理和质量控制。该数据集适用于深度学习模型训练，以解决心脏磁共振成像中的加速重建问题。

The CMR×Recon dataset was jointly developed by the Human Phenome Institute of Fudan University and other institutions. It encompasses multi-contrast, multi-view, multi-slice, and multi-coil cardiac magnetic resonance imaging (CMR) data acquired from 300 healthy volunteers. This dataset includes cardiac cine and mapping sequences, and provides fine-grained segmentations of the myocardium and ventricles for all subjects. The core goal of this dataset is to promote the advancement of cardiac MRI reconstruction technologies by offering standardized evaluation benchmarks and open public access. Data collection was conducted using a 3T MRI scanner and a 32-channel cardiac coil, followed by data processing and quality control via specialized software. This dataset is suitable for training deep learning models to solve accelerated reconstruction problems in cardiac magnetic resonance imaging.