High-Fidelity Functional Ultrasound Reconstruction via A Visual Auto-Regressive Framework

TL;DR

This paper introduces a novel visual auto-regressive framework to improve high-fidelity functional ultrasound reconstruction, addressing data scarcity and signal degradation issues in neurovascular imaging.

Contribution

It proposes a new auto-regressive approach that enhances ultrasound image quality despite limited data and signal challenges.

Findings

Improved reconstruction quality demonstrated in experiments

Enhanced robustness to data scarcity and signal degradation

Potential for broader clinical application

Abstract

Functional ultrasound (fUS) imaging provides exceptional spatiotemporal resolution for neurovascular mapping, yet its practical application is significantly hampered by critical challenges. Foremost among these are data scarcity, arising from ethical considerations and signal degradation through the cranium, which collectively limit dataset diversity and compromise the fairness of downstream machine learning models.