Fast Multi-Stack Slice-to-Volume Reconstruction via Multi-Scale Unrolled Optimization

TL;DR

The paper introduces a fast, non-rigid multi-stack slice-to-volume reconstruction framework for fetal brain MRI that fuses three orthogonal slice stacks and refines slice poses through multi-scale unrolled optimization integrated with model-based reconstruction, achieving sub-10-second total reconstruction and ~1-second pose estimation. By parameterizing non-rigid motion with displacement fields and employing a learned residual flow across resolutions, the method delivers accurate volumes while significantly reducing runtime compared to state-of-the-art INR and optimization-based SVR approaches. Evaluations on simulated and FeTA clinical data demonstrate robustness to rotations, translations, and noise, with performance comparable to SVoRT and NeSVoR while offering substantial speedups. The approach generalizes to non-rigid SVR tasks (e.g., placental MRI) and paves the way for real-time, scanner-side feedback during MRI acquisition, potentially guiding acquisition planning in real time.

Abstract

Fully convolutional networks have become the backbone of modern medical imaging due to their ability to learn multi-scale representations and perform end-to-end inference. Yet their potential for slice-to-volume reconstruction (SVR), the task of jointly estimating 3D anatomy and slice poses from misaligned 2D acquisitions, remains underexplored. We introduce a fast convolutional framework that fuses multiple orthogonal 2D slice stacks to recover coherent 3D structure and refines slice alignment through lightweight model-based optimization. Applied to fetal brain MRI, our approach reconstructs high-quality 3D volumes in under 10s, with 1s slice registration and accuracy on par with state-of-the-art iterative SVR pipelines, offering more than speedup. The framework uses non-rigid displacement fields to represent transformations, generalizing to other SVR problems like fetal body and placental MRI. Additionally, the fast inference time paves the way for real-time, scanner-side volumetric feedback during MRI acquisition.

Figures (12)

• Figure 1: Fast Multi-Stack Slice-to-Volume Reconstruction. Our proposed multi-stack SVR framework takes as input three motion-corrupted stacks of 2D slices and reconstructs a volume (1 second). Super-resolution is performed With optional optimization (7 seconds).
• Figure 2: Method Overview. (A) SVR pipeline combines convolutional pose estimation with model-based reconstruction. (B) Iterative 2D+3D blocks refine slice pose estimates at resolution $s$ through simulated slice generation and flow field updates.
• Figure 3: Clinical Evaluation. Reconstructions for clinical subjects (GA 20–35 weeks) for all methods: SVoRT, SVoRT + NeSVoR, cSVR, cSVR + Refine, cSVR + NeSVoR. Our proposed fast method, cSVR + Refine, achieves high-quality reconstructions comparable to state of the art, with high grey and white matter contrast (green arrows). Our method as well as SVoRT struggles in cases of image corruption as seen by the red arrows, where the reconstructions fail to exclude noisy areas of a slice.
• Figure 4: Performance evaluation and sensitivity analysis. Robustness across methods to input stack perturbations (translation, rotation, noise), evaluated via registration accuracy (top) and reconstruction quality (bottom). Our method, cSVR + refine, is robust across high levels of translation and rotation and our method coupled with NeSVoR reconstruction achieves the best overall performance.
• Figure 5: Inference Time of Registration Methods Time to predict slice poses on clinical subjects of varying input sizes, comparing SVoRT vs cSVR.