PrIINeR: Towards Prior-Informed Implicit Neural Representations for Accelerated MRI

2025-08-11

AI Summary

Accelerated MRI reconstruction at high acceleration factors often suffers from structural loss and aliasing artifacts, primarily due to insufficient anatomical priors in existing implicit neural representation (INR) methods. To address this, we propose a prior-guided INR framework that embeds anatomical priors from a pre-trained deep model into the INR architecture. Our approach jointly leverages population-level prior modeling and instance-level optimization, and introduces a dual k-space data consistency constraint to synergistically integrate deep priors with the physical forward model. The entire framework is trained end-to-end, balancing generalizability and reconstruction fidelity. Evaluated on the NYU fastMRI dataset, our method significantly outperforms state-of-the-art INR and conventional learning-based baselines. At an acceleration factor of ×8, it effectively suppresses artifacts while preserving fine anatomical details, achieving superior quantitative metrics (PSNR/SSIM) and enhanced clinical interpretability.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Representation Learning for Vision
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Accelerating Magnetic Resonance Imaging (MRI) reduces scan time but often degrades image quality. While Implicit Neural Representations (INRs) show promise for MRI reconstruction, they struggle at high acceleration factors due to weak prior constraints, leading to structural loss and aliasing artefacts. To address this, we propose PrIINeR, an INR-based MRI reconstruction method that integrates prior knowledge from pre-trained deep learning models into the INR framework. By combining population-level knowledge with instance-based optimization and enforcing dual data consistency, PrIINeR aligns both with the acquired k-space data and the prior-informed reconstruction. Evaluated on the NYU fastMRI dataset, our method not only outperforms state-of-the-art INR-based approaches but also improves upon several learning-based state-of-the-art methods, significantly improving structural preservation and fidelity while effectively removing aliasing artefacts.PrIINeR bridges deep learning and INR-based techniques, offering a more reliable solution for high-quality, accelerated MRI reconstruction. The code is publicly available on https://github.com/multimodallearning/PrIINeR.

Problem

• Improve MRI image quality at high acceleration factors
• Integrate prior knowledge into neural representations for MRI
• Reduce structural loss and aliasing artefacts in MRI

Innovation

• Integrates prior knowledge into INR framework
• Enforces dual data consistency for alignment
• Combines population-level and instance-based optimization