MoE-dqINR: A Unified Mixture-of-Experts Implicit Neural Representation Framework for Scan-Specific Dynamic and Quantitative MRI Reconstruction

• URL: http://arxiv.org/abs/2605.31302v1
• Date: Fri, 29 May 2026 13:36:25 GMT
• Title: MoE-dqINR: A Unified Mixture-of-Experts Implicit Neural Representation Framework for Scan-Specific Dynamic and Quantitative MRI Reconstruction
• Abstract summary: Underd magnetic resonance imaging (MRI) reconstruction seeks to recover temporally or contrast-varying image series from incomplete multicoil k-space data.<n>Existing scan-specific implicit neural representations (INRs) often use monolithic coordinate fields, explicit subspaces, motion or deformation models, or sequence-specific quantitative signal models.<n>We propose MoE-dqINR, a scan-specific multicoil MRI reconstruction framework that factorizes the image-domain representation into shared spatial experts and a state-conditioned routing pathway.
• Score: 10.062510791924684
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Undersampled magnetic resonance imaging (MRI) reconstruction seeks to recover temporally or contrast-varying image series from incomplete multicoil k-space data while preserving state-dependent fidelity for dynamic and quantitative MRI (qMRI). Existing scan-specific implicit neural representations (INRs) often use monolithic spatiotemporal coordinate fields, explicit subspaces, motion or deformation models, calibration variables, or sequence-specific quantitative signal models. These design choices can limit flexibility in sharing spatial information while adapting image synthesis across acquisition states. Moreover, many INR-based baselines remain computationally demanding, typically requiring per-scan optimization times on the order of hundreds to thousands of seconds. We propose MoE-dqINR, a scan-specific multicoil MRI reconstruction framework that factorizes the image-domain representation into shared spatial experts and a state-conditioned routing pathway. Spatial experts encode reusable coordinate-dependent image content, whereas routing weights, conditioned on ordered acquisition states, synthesize each dynamic frame or contrast state from a common expert bank. The representation is coupled to a multicoil MRI forward model, uses the normalized state index to drive routing in both dynamic and quantitative MRI. By separating shared spatial representation from state-dependent synthesis, the framework provides an image-first architecture for dynamic and quantitative MRI while reducing scan-specific INR optimization to approximately 30 s per scan in our experiments. The proposed formulation establishes state-conditioned mixture-of-experts INR as a scan-specific multicoil MRI reconstruction prior that unifies shared spatial representation, dynamic- and qMRI-specific synthesis, and practical per-scan efficiency.

Related papers

• Mamba-driven MRI-to-CT Synthesis for MRI-only Radiotherapy Planning [4.3531418255555545]
  Mamba is a state-space model for cross-modality translation.<n>We adapt both the U-Mamba and the SegMamba architecture, originally proposed for segmentation, to perform cross-modality image generation.<n>Our 3D Mamba architecture effectively captures complex features and long-range volumetric dependencies, thus allowing accurate CT synthesis.
  arXiv  Detail & Related papers  (2026-03-24T14:56:39Z)
• MRI2Qmap: multi-parametric quantitative mapping with MRI-driven denoising priors [0.0280162601599316]
  We introduce MRI2Qmap, a plug-and-play quantitative reconstruction framework.<n>It integrates the physical acquisition model with priors learned from deep denoising autoencoders pretrained on large multimodal weighted-MRI datasets.<n>The proposed method is validated on highly accelerated 3D whole-brain MRF data from both in-vivo and simulated acquisitions.
  arXiv  Detail & Related papers  (2026-03-11T21:23:52Z)
• A Unified Model for Compressed Sensing MRI Across Undersampling Patterns [69.19631302047569]
  We propose a unified MRI reconstruction model robust to various measurement undersampling patterns and image resolutions.<n>Our model improves SSIM by 11% and PSNR by 4 dB over a state-of-the-art CNN (End-to-End VarNet) with 600$times$ faster inference than diffusion methods.
  arXiv  Detail & Related papers  (2024-10-05T20:03:57Z)
• Towards General Text-guided Image Synthesis for Customized Multimodal Brain MRI Generation [51.28453192441364]
  Multimodal brain magnetic resonance (MR) imaging is indispensable in neuroscience and neurology. Current MR image synthesis approaches are typically trained on independent datasets for specific tasks. We present TUMSyn, a Text-guided Universal MR image Synthesis model, which can flexibly generate brain MR images.
  arXiv  Detail & Related papers  (2024-09-25T11:14:47Z)
• NeuroPictor: Refining fMRI-to-Image Reconstruction via Multi-individual Pretraining and Multi-level Modulation [55.51412454263856]
  This paper proposes to directly modulate the generation process of diffusion models using fMRI signals. By training with about 67,000 fMRI-image pairs from various individuals, our model enjoys superior fMRI-to-image decoding capacity.
  arXiv  Detail & Related papers  (2024-03-27T02:42:52Z)
• Volumetric Reconstruction Resolves Off-Resonance Artifacts in Static and Dynamic PROPELLER MRI [76.60362295758596]
  Off-resonance artifacts in magnetic resonance imaging (MRI) are visual distortions that occur when the actual resonant frequencies of spins within the imaging volume differ from the expected frequencies used to encode spatial information. We propose to resolve these artifacts by lifting the 2D MRI reconstruction problem to 3D, introducing an additional "spectral" dimension to model this off-resonance.
  arXiv  Detail & Related papers  (2023-11-22T05:44:51Z)
• Robust Depth Linear Error Decomposition with Double Total Variation and Nuclear Norm for Dynamic MRI Reconstruction [15.444386058967579]
  There are still problems with dynamic MRI k-space reconstruction based on Compressed Sensing (CS) In this paper, we propose a novel robust lowrank dynamic MRI reconstruction optimization model via highly under-sampled Fourier Transform (DFT) Experiments on dynamic MRI data demonstrate the superior performance proposed method in terms of both reconstruction accuracy and time complexity.
  arXiv  Detail & Related papers  (2023-10-23T13:34:59Z)
• CMRxRecon: An open cardiac MRI dataset for the competition of accelerated image reconstruction [62.61209705638161]
  There has been growing interest in deep learning-based CMR imaging algorithms. Deep learning methods require large training datasets. This dataset includes multi-contrast, multi-view, multi-slice and multi-coil CMR imaging data from 300 subjects.
  arXiv  Detail & Related papers  (2023-09-19T15:14:42Z)
• Efficient MRI Parallel Imaging Reconstruction by K-Space Rendering via Generalized Implicit Neural Representation [16.63720411275398]
  This study presents a generalized implicit neural representation (INR)-based framework for MRI PI reconstruction.<n>The proposed method overcomes limitations by leveraging prior knowledge of voxel-specific features.<n>Experiments on publicly available MRI datasets demonstrate the superior performance of the proposed method.
  arXiv  Detail & Related papers  (2023-09-12T09:07:03Z)
• CoNeS: Conditional neural fields with shift modulation for multi-sequence MRI translation [5.662694302758443]
  Multi-sequence magnetic resonance imaging (MRI) has found wide applications in both modern clinical studies and deep learning research. It frequently occurs that one or more of the MRI sequences are missing due to different image acquisition protocols or contrast agent contraindications of patients. One promising approach is to leverage generative models to synthesize the missing sequences, which can serve as a surrogate acquisition.
  arXiv  Detail & Related papers  (2023-09-06T19:01:58Z)
• Multi-Modal MRI Reconstruction with Spatial Alignment Network [51.74078260367654]
  In clinical practice, magnetic resonance imaging (MRI) with multiple contrasts is usually acquired in a single study. Recent researches demonstrate that, considering the redundancy between different contrasts or modalities, a target MRI modality under-sampled in the k-space can be better reconstructed with the helps from a fully-sampled sequence. In this paper, we integrate the spatial alignment network with reconstruction, to improve the quality of the reconstructed target modality.
  arXiv  Detail & Related papers  (2021-08-12T08:46:35Z)
• Adaptive Gradient Balancing for UndersampledMRI Reconstruction and Image-to-Image Translation [60.663499381212425]
  We enhance the image quality by using a Wasserstein Generative Adversarial Network combined with a novel Adaptive Gradient Balancing technique. In MRI, our method minimizes artifacts, while maintaining a high-quality reconstruction that produces sharper images than other techniques.
  arXiv  Detail & Related papers  (2021-04-05T13:05:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.