arXiv 2603.044382026-03eess.IVcs.AIcs.LG

CogGen: 认知负荷启发的全无监督深度生成模型用于压缩感知MRI重建

CogGen: Cognitive-Load-Inspired Fully Unsupervised Deep Generative Modeling for Compressively Sampled MRI Reconstruction

Medical AI Research Centre, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（深圳先进技术研究院医学人工智能研究中心）；School of Biomedical Engineering, Southern Medical University（南方医科大学生物医学工程学院）；School of Mathematics and Statistics, Xi’an Jiaotong University（西安交通大学数学与统计学院）；Paul C. Lauterbur Research Centre for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（深圳先进技术研究院Paul C. Lauterbur生物医学成像研究中心）

Qingyong Zhu, Yumin Tan, Xiang Gu, Dong Liang

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出CogGen框架，基于认知易到难原则，通过自定进度课程学习和MRI感知双阈值加权策略，将CS-MRI重建分解为分阶段反演问题，理论证明降低局部充分迭代界和累积噪声放大界，实验优于现有无监督和有监督方法。

摘要：

全无监督深度生成建模（FU-DGM）为压缩感知磁共振成像（CS-MRI）重建提供了巨大潜力。代表性的FU-DGM公式，如深度图像先验（DIP）和隐式神经表示（INR），利用架构偏置在图像空间中诱导与正向观测对齐的低维流形。然而，由于底层逆系统高度病态，FU-DGM中长时间的迭代拟合通常导致效率低下和噪声放大。本文受认知易到难学习原则的启发，提出CogGen，一种将CS-MRI重建重新表述为分阶段反演问题的FU-DGM框架。具体地，CogGen通过MRI感知的双阈值加权准则实现自定进度课程学习（SPCL）驱动的渐进调度策略，该准则自适应地调节k空间测量参与。数据一致性残差阈值评估当前生成器的拟合可靠性，而k空间半径阈值控制阶段性的测量暴露，从而避免整个优化过程中的均匀拟合。理论上，我们的分析表明，当早期阶段倾向于易拟合的测量时，CogGen产生更低的局部充分迭代界和更小的累积噪声放大界，解释了CogGen在有限迭代预算内改进的收敛行为和重建保真度。数值实验表明，CogGen的两种实例化，CogGen-DIP和CogGen-INR，在包括无监督和有监督流程在内的现有CS-MRI重建技术中实现了优越的性能。

英文摘要：

Fully unsupervised deep generative modeling (FU-DGM) offers significant potential for compressively sampled magnetic resonance imaging (CS-MRI) reconstruction. Representative FU-DGM formulations, such as deep image prior (DIP) and implicit neural representation (INR), employ architectural bias to induce a low-dimensional manifold in the image space that aligns with the forward observation. However, as the underlying inverse system is highly ill-posed, prolonged iterative fitting in FU-DGM typically leads to poor efficiency and noise amplification. In this paper, guided by the cognitive principle of easy-to-hard learning, we propose CogGen, an FU-DGM framework that reformulates CS-MRI reconstruction as a staged inversion problem. Specifically, CogGen implements an self-paced curriculum learning (SPCL)-driven progressive scheduling strategy through an MRI-aware dual-threshold weighting criterion, which adaptively regulates k-space measurement participation. The data-consistency residual thresholding evaluates the fitting reliability of the current generator, while the k-space radius thresholding controls stage-wise measurement exposure, thereby avoiding uniform fitting throughout optimization. Theoretically, our analysis shows that, when early stages favor easy-to-fit measurements, CogGen yields a reduced local sufficient-iteration bound and a smaller cumulative noise-amplification bound, explaining the improved convergence behavior and reconstruction fidelity of CogGen within a finite iteration budget. Numerical experiments demonstrate that both CogGen instantiations, CogGen-DIP and CogGen-INR, achieve superior performance over prevailing CS-MRI reconstruction techniques, including unsupervised and supervised pipelines.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。