FSC-loss: A Frequency-domain Structure Consistency Learning Approach for Signal Data Recovery and Reconstruction

2025-01-08

AI Summary

Magnetic particle imaging (MPI) faces a critical challenge in super-resolution reconstruction of the signal matrix (SM)—i.e., rapidly recovering high-resolution SM from highly undersampled, low-fidelity measurements to drastically reduce acquisition time. Method: We propose a Transformer-based frequency-domain structural-aware reconstruction framework. It introduces a novel frequency-domain structural consistency loss and a data-component embedding strategy to explicitly model both global and local SM structures in the frequency domain, thereby overcoming the bottleneck in high-frequency information recovery. The method integrates frequency-domain analysis, structural priors, and multi-scale feature embedding. Contribution/Results: Our approach achieves state-of-the-art performance on both public and simulated MPI datasets: it reconstructs high-resolution SM at only 1/16 sampling rate in under 15 seconds (nRMSE = 0.041), yielding >60× acceleration. It has been successfully deployed on three in-house MPI systems, significantly enhancing clinical imaging performance.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Machine Learning: Matrix & Tensor Methods
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

A core challenge for signal data recovery is to model the distribution of signal matrix (SM) data based on measured low-quality data in biomedical engineering of magnetic particle imaging (MPI). For acquiring the high-resolution (high-quality) SM, the number of meticulous measurements at numerous positions in the field-of-view proves time-consuming (measurement of a 37x37x37 SM takes about 32 hours). To improve reconstructed signal quality and shorten SM measurement time, existing methods explore to generating high-resolution SM based on time-saving measured low-resolution SM (a 9x9x9 SM just takes about 0.5 hours). However, previous methods show poor performance for high-frequency signal recovery in SM. To achieve a high-resolution SM recovery and shorten its acquisition time, we propose a frequency-domain structure consistency loss function and data component embedding strategy to model global and local structural information of SM. We adopt a transformer-based network to evaluate this function and the strategy. We evaluate our methods and state-of-the-art (SOTA) methods on the two simulation datasets and four public measured SMs in Open MPI Data. The results show that our method outperforms the SOTA methods in high-frequency structural signal recovery. Additionally, our method can recover a high-resolution SM with clear high-frequency structure based on a down-sampling factor of 16 less than 15 seconds, which accelerates the acquisition time over 60 times faster than the measurement-based HR SM with the minimum error (nRMSE=0.041). Moreover, our method is applied in our three in-house MPI systems, and boost their performance for signal reconstruction.

Problem

• Magnetic Particle Imaging
• High Resolution Signal Generation
• Measurement Time Reduction

Innovation

• FSC-loss
• High-frequency signal recovery
• Magnetic Particle Imaging (MPI)