SUFFICIENT: A scan-specific unsupervised deep learning framework for high-resolution 3D isotropic fetal brain MRI reconstruction

2025-05-23

AI Summary

Severe motion artifacts in 2D fetal brain MRI slices hinder high-fidelity, isotropic 3D reconstruction. Method: We propose a training-free, high-resolution isotropic 3D reconstruction framework tailored to individual scan characteristics. It formulates an unsupervised joint optimization that concurrently solves slice-to-volume registration (SVR) and super-resolution reconstruction (SRR). The method integrates differentiable rigid registration, deep image prior (DIP)-guided high-resolution volume synthesis, and a differentiable forward degradation model, enabling end-to-end optimization without external supervision. Contribution/Results: Our approach significantly improves reconstruction quality on both simulated and real clinical fetal MRI data, yielding sub-millimeter, high-fidelity, isotropic 3D fetal brain volumes. Crucially, it achieves the first SVR–SRR co-optimization without relying on large-scale annotated fetal MRI datasets—demonstrating robustness, generalizability, and clinical practicality.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

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
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

High-quality 3D fetal brain MRI reconstruction from motion-corrupted 2D slices is crucial for clinical diagnosis. Reliable slice-to-volume registration (SVR)-based motion correction and super-resolution reconstruction (SRR) methods are essential. Deep learning (DL) has demonstrated potential in enhancing SVR and SRR when compared to conventional methods. However, it requires large-scale external training datasets, which are difficult to obtain for clinical fetal MRI. To address this issue, we propose an unsupervised iterative SVR-SRR framework for isotropic HR volume reconstruction. Specifically, SVR is formulated as a function mapping a 2D slice and a 3D target volume to a rigid transformation matrix, which aligns the slice to the underlying location in the target volume. The function is parameterized by a convolutional neural network, which is trained by minimizing the difference between the volume slicing at the predicted position and the input slice. In SRR, a decoding network embedded within a deep image prior framework is incorporated with a comprehensive image degradation model to produce the high-resolution (HR) volume. The deep image prior framework offers a local consistency prior to guide the reconstruction of HR volumes. By performing a forward degradation model, the HR volume is optimized by minimizing loss between predicted slices and the observed slices. Comprehensive experiments conducted on large-magnitude motion-corrupted simulation data and clinical data demonstrate the superior performance of the proposed framework over state-of-the-art fetal brain reconstruction frameworks.

Problem

• Reconstructing high-resolution 3D fetal brain MRI from motion-corrupted 2D slices
• Overcoming need for large external datasets in deep learning-based MRI reconstruction
• Improving slice-to-volume registration and super-resolution reconstruction accuracy

Innovation

• Unsupervised deep learning for 3D MRI reconstruction
• CNN-based slice-to-volume registration alignment
• Deep image prior framework for super-resolution