Problem
Research questions and friction points this paper is trying to address.
Magnetic Resonance Imaging
Image Reconstruction
Medical Imaging
Adaptive Local Neighborhood-Based Neural Networks for MR Image Reconstruction From Undersampled Data
🤖 AI Summary
To address poor generalizability and weak adaptability to diverse scanning protocols in undersampled MRI reconstruction, this paper proposes a novel Reconstruction-time Online Local Modeling (ROLM) paradigm. During inference, ROLM dynamically retrieves semantically similar training samples to construct a local neighborhood and instantaneously fits a lightweight local neural network for scan-specific reconstruction. This approach eliminates reliance on a fixed global model, enabling seamless adaptation to heterogeneous acquisition protocols and incremental updates to the training set. Integrated within a deep unrolling framework and combined with 1D variable-density random undersampling, ROLM achieves significant PSNR/SSIM improvements over both conventional global models and state-of-the-art adaptive methods at 4× and 8× acceleration factors. Reconstructed images meet clinical usability standards, striking an optimal balance between high fidelity and low latency.
📝 Abstract
Recent medical image reconstruction techniques focus on generating high-quality medical images suitable for clinical use at the lowest possible cost and with the fewest possible adverse effects on patients. Recent works have shown significant promise for reconstructing MR images from sparsely sampled k-space data using deep learning. In this work, we propose a technique that rapidly estimates deep neural networks directly at reconstruction time by fitting them on small adaptively estimated neighborhoods of a training set. In brief, our algorithm alternates between searching for neighbors in a data set that are similar to the test reconstruction, and training a local network on these neighbors followed by updating the test reconstruction. Because our reconstruction model is learned on a dataset that is in some sense similar to the image being reconstructed rather than being fit on a large, diverse training set, it is more adaptive to new scans. It can also handle changes in training sets and flexible scan settings, while being relatively fast. Our approach, dubbed LONDN-MRI, was validated on multiple data sets using deep unrolled reconstruction networks. Reconstructions were performed at four fold and eight fold undersampling of k-space with 1D variable-density random phase-encode undersampling masks. Our results demonstrate that our proposed locally-trained method produces higher-quality reconstructions compared to models trained globally on larger datasets as well as other scan-adaptive methods.
Innovation
Methods, ideas, or system contributions that make the work stand out.
LONDN-MRI
self-adjusting local neural networks
high-quality image reconstruction
🔎 Similar Papers
No similar papers found.
