Data for: "A Neural-Network-Based Convex Regularizer for Inverse Problems"

Data for: "A Neural-Network-Based Convex Regularizer for Inverse Problems". The corresponding scripts can be accessed on GitHub (https://github.com/axgoujon/convex_ridge_regularizers). The data is organized as follows: - ct_data_sets.tar.gz: contains preprocessed validation (aka calibration) and test sets with: ground truth images, FBP reconstructions, measurements, for the various settings explored (3 noise levels). - mri_data_sets.tar.gz: contains preprocessed validation (aka calibration) and test sets with: subsampling cartesian masks, sensitivity masks, ground truth image, measurements, for the various settings explored: single- and multi-coil MRI, various acceleration rates (2, 4, and 8), synthetic noise, and different image types (fat suppression or not). For completeness, the code used to generate the preprocessed data from the raw data can be found on the GitHub repository.本数据集对应论文《基于神经网络的逆问题凸正则化器》，相关配套脚本可于GitHub仓库（https://github.com/axgoujon/convex_ridge_regularizers）获取。 数据集结构如下： - ct_data_sets.tar.gz：包含预处理后的验证集（亦称校准集）与测试集，涵盖以下内容： 真值图像、 滤波反投影（Filtered BackProjection, FBP）重建结果、 测量数据，覆盖本次实验探索的多种设置（包含3种噪声水平）。 - mri_data_sets.tar.gz：包含预处理后的验证集（亦称校准集）与测试集，涵盖以下内容： 笛卡尔下采样掩码、 灵敏度掩码、 真值图像、 测量数据，覆盖本次实验探索的多种设置：单线圈与多线圈磁共振成像（Magnetic Resonance Imaging, MRI）、不同加速倍率（2、4、8倍）、合成噪声，以及不同图像类型（是否抑脂）。 为保证完整性，从原始数据生成预处理数据集的相关代码亦可在该GitHub仓库中获取。