Simulated nanoCT Datasets

Simulated parallel-beam and fan-beam nano-CT datasets with per-angle random phantom shifts and rotations. The dataset generation is described in the paper Learning-based approaches for reconstructions with inexact operators in nanoCT applications. Code to use the data is available at https://gitlab.informatik.uni-bremen.de/inn4ip/cond-inn4nanoct (in particular the module src/learned_reco/data.py).本数据集为包含逐角度随机体模位移与旋转的模拟平行束及扇束纳米计算机断层扫描（nano-CT）数据集。该数据集的生成方法已在论文《纳米CT应用中基于不准确算子的重建学习方法》（Learning-based approaches for reconstructions with inexact operators in nanoCT applications）中详述。该数据集的使用代码可从https://gitlab.informatik.uni-bremen.de/inn4ip/cond-inn4nanoct 获取，具体对应src/learned_reco/data.py模块。