Self-Supervised Isotropic MRI Volume Restoration from Complementary Contrast-Plane Acquisition using Two-Phase Fast Score-based Models

Motivation: Self-supervised isotropic volume reconstruction is essential to address the limitations of anisotropic multi-contrast MRI, where differing planar orientations hinder diagnostic pooling. Goal(s): This study aims to generate isotropic, multi-contrast MR volumes from anisotropic data to unify diagnostic information across contrasts. Approach: A self-supervised score-based framework trains on anisotropic images to iteratively refine volumetric estimates across contrasts. Results: Demonstrating superior image quality on brain MRI datasets, the method advances the usability of existing anisotropic multi-contrast protocols in clinical practice. Impact: This method enhances clinical MRI protocols by enabling isotropic multi-contrast volume reconstruction from anisotropic data, improving diagnostic consistency across contrasts. It reduces the need for extended scan times, maximizing data utility and facilitating broader clinical insights in routine practice.