Datasets based on the X-space method in Magnetic Particle Imaging

Citation Author(s):
Submitted by:
Yaxin Shang
Last updated:
Tue, 07/11/2023 - 05:34
Data Format:
0 ratings - Please login to submit your rating.


This simulation dataset contains five types of data: resolutions, vessels, vessel stenosis, tumors, and shape combinations. There are a total of 1000 original binary images. Besides, we set different gray values on images with multiple connected domains to simulate different concentration of magnetic nanoparticles. Next, the images are subjected to operations such as image inversion and image rotation. The final dataset contains 20,000 images. we applied the X-space method based on the X-space theory and we generated the simulated image of magnetic particle imaging. This dataset serves as a valuable resource for benchmarking and comparing different methods in the field of MPI image reconstruction.


This dataset consists of native images and phantoms. The native images are generated using the X-space method, while the phantoms represent the true distribution of MNPs concentration in the form of original images. There are 20,000 instances of both native images and phantoms, and each pair has a one-to-one correspondence.