跳到主要导航 跳到搜索 跳到主要内容

Deliver-MPI: Diffusion Embedded-Physics Model for Magnetic Particle Imaging Reconstruction under Strong Liver Background

  • Ziwei Chen
  • , Wenxuan Zou
  • , Gen Shi
  • , Jian'an Ye
  • , Ning Sun
  • , Zhongwei Bian
  • , Yu An*
  • , Jie Tian*
  • *此作品的通讯作者
  • Beihang University

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

Magnetic Particle Imaging (MPI) is an emerging, radiationfree modality for quantitatively mapping superparamagnetic tracer concentration, yet liver uptake often overwhelms nearby tumor signals. We propose a physics-constrained diffusion reconstruction that injects the system matrix into a pre-trained diffusion prior, alternating denoising with a onestep data-consistency update and null-space conditioning. Without changing hardware or tracers, the method enforces physical fidelity while preserving fine structures. On simulated and measured liver-tumor phantoms, it reduces artifacts and more reliably recovers small lesions adjacent to strong liver signals compared with traditional reconstructions.

源语言英语
主期刊名ISBI 2026 - 23rd IEEE International Symposium on Biomedical Imaging
出版商IEEE Computer Society
ISBN(电子版)9798331577636
DOI
出版状态已出版 - 2026
活动23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026 - London, 英国
期限: 8 4月 202611 4月 2026

出版系列

姓名Proceedings - International Symposium on Biomedical Imaging
2026-April
ISSN(印刷版)1945-7928
ISSN(电子版)1945-8452

会议

会议23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026
国家/地区英国
London
时期8/04/2611/04/26

指纹

探究 'Deliver-MPI: Diffusion Embedded-Physics Model for Magnetic Particle Imaging Reconstruction under Strong Liver Background' 的科研主题。它们共同构成独一无二的指纹。

引用此