MPI-MNIST
收藏资源简介:
MPI-MNIST数据集是基于MNIST手写数字数据集构建的,用于模拟磁粒子成像(MPI)测量。该数据集由不来梅大学工业数学中心创建,包含基于模型系统矩阵生成的模拟MPI测量数据,并结合了来自临床前MPI扫描仪设备的真实噪声测量。数据集提供了一个真实且灵活的算法测试环境,旨在解决MPI图像重建中的噪声模型问题。通过该数据集,研究人员可以验证和改进MPI重建算法,特别是在处理复杂噪声分布时的表现。
The MPI-MNIST dataset is built upon the MNIST handwritten digit dataset for simulating magnetic particle imaging (MPI) measurements. Created by the Center for Industrial Mathematics at the University of Bremen, this dataset includes simulated MPI measurement data generated based on a model system matrix, combined with real noise measurements from preclinical MPI scanner devices. The dataset provides a realistic and flexible algorithmic test environment, aiming to address the noise modeling problem in MPI image reconstruction. Through this dataset, researchers can validate and improve MPI reconstruction algorithms, particularly their performance when handling complex noise distributions.




