Anisotropic edge-preserving network for resolution enhancement in unidirectional Cartesian magnetic particle imaging

Abstract

Objective: Magnetic particle imaging (MPI) is a novel imaging modality. It is crucial to acquire accurate localization of the superparamagnetic iron oxide (SPIO) nanoparticles distributions in MPI. However, the spatial resolution of unidirectional Cartesian trajectory MPI exhibits anisotropy, which blurs the boundaries of MPI images and makes precise localization difficult. In this paper, we propose an anisotropic edge-preserving network (AEP-net) to alleviate the anisotropic resolution of MPI. Methods: AEP-net resolve the resolution anisotropy by constructing an asymmertic convolution. To recover the edge information, we design the uncertainty region module. In addition, we evaluated the performance of the proposed AEP-net model by using simulations and experimental data. Results: The results show that the AEP-net model alleviates the anisotropy of the unidirectional Cartesian trajectory and preserves edge details in the MPI image. By comparing the visualization results and the metrics, we demonstrate that our method is superior to other methods. Significance: The proposed method produces accurate visualization in unidirectional Cartesian devices and promotes accurate quantization, which promote the biomedical applications using MPI.

Combination of time domain-system matrix and x-space methods to reconstruct magnetic particle images with isotropic resolution

“…We consider that imaging SPIONs is, after recovery of the fundamental frequency signal, an LSI process (Lu et al 2013), and that x-space reconstruction of particle images involves convolution of particle distribution and the PSF kernel (Goodwill and Conolly 2011). Within the unidirectional Cartesian trajectory applied, the magnetic field slowly changes perpendicularly to the excitation, which makes the PSF anisotropic (Shang et al 2023). If we construct a forward model to describe the relationship between the particle distribution and the image reconstructed using x-space, which also involves convolution of the PSF kernel, then we can use an inverse problem solver to obtain an image without relying on the PSF.…”

“…The primary cause of resolution anisotropy in x-space reconstructed images is the unidirectional Cartesian trajectory of the scanning path (Shang et al 2023). This issue is also evident in the reconstruction outcomes of the Cartesian trajectory-based system matrix method, with limited resolution enhancement (Knopp et al 2009).…”

“…The two resulting anisotropic images are merged into one isotropic image. Recently, a deep learning method was proposed to reduce the anisotropic resolution of the unidirectional x-space reconstructed image (Shang et al 2023).…”

BSS-TFNet: Attention-Enhanced Background Signal Suppression Network for Time-Frequency Spectrum in Magnetic Particle Imaging

“…And there are other works that proposed novel deep learning frameworks for SM super-resolution [34], [35]. Besides, deep learning has been adopted as a promising approach in other MPI tasks, including the resolution improvement of MPI images by CNN [36], [37], view imputation in projection MPI imaging using a Generative adversarial network [38] and crossdomain knowledge transfer strategy [39], and MPI image reconstruction [40]. For the denoising tasks of MPI, Peng et al [41] proposed a multi-scale dual domain network, which combined time domain and frequency domain information to remove the background signals and filter the MPI signals.…”