FFL-MPI-Data (v.1.0): a public projection imaging dataset for field free line magnetic particle imaging

This repository comprises a public projection imaging dataset named FFL-MPI-Data for field free line magnetic particle imaging.FFL-MPI-Data contains the following datasets:measured dataset: using a homemade scanner with projection imaging sequence, collecting raw voltage signals, sinograms, and reconstructed images from 100 physical phantoms.simulated dataset: generated by establishing a simulation framework that replicates the key parameters of the experimental scanner, with MNIST dataset as digital phantoms. For deep learning applications, the FFL-MPI-Data simulation dataset supports model pre-training, while the measured dataset is suitable for model fine-tuning. Given the limited scale of the measured dataset, data augmentation techniques can be applied to further expand its size. Additionally, FFL-MPI-Data provides raw voltage signals from experimental measurements. Although only the 2nd to 7th harmonics have been extracted in this study, other intermodulation harmonics can be further analyzed to facilitate image reconstruction algorithm development.本仓库包含一款面向无场线磁粒子成像（field free line magnetic particle imaging）的公开投影成像数据集，命名为FFL-MPI-Data。FFL-MPI-Data包含以下两类数据集： 实测数据集：采用自研扫描仪搭配投影成像序列，采集了100个体模的原始电压信号、正弦图（sinograms）以及重建图像。 仿真数据集：通过搭建与实验扫描仪关键参数一致的仿真框架生成，以MNIST数据集作为数字体模。 针对深度学习应用场景，FFL-MPI-Data的仿真数据集可用于模型预训练，实测数据集则适用于模型微调。考虑到实测数据集规模有限，可通过数据增强技术进一步扩充其样本量。此外，FFL-MPI-Data还提供了实验测量获取的原始电压信号。尽管本研究仅提取了2至7次谐波，但可进一步分析其他互调谐波，以助力图像重建算法的研发。