Super-resolution Ultrasound Localization Microscopy through Deep Learning

TL;DR

This paper introduces Deep Ultrasound Localization Microscopy (Deep-ULM), a deep learning-based method that achieves super-resolution vascular imaging from high-density ultrasound data, significantly reducing acquisition time and enabling real-time in-vivo imaging.

Contribution

The work presents a novel deep learning framework for super-resolution ultrasound microscopy that effectively handles high microbubble densities, improving speed and accuracy over traditional methods.

Findings

Achieves super-resolution in dense microbubble scenarios.

Operates in real-time, processing up to 1250 patches per second with GPU.

Validates effectiveness both in-silico and in-vivo experiments.

Abstract

Ultrasound localization microscopy has enabled super-resolution vascular imaging through precise localization of individual ultrasound contrast agents (microbubbles) across numerous imaging frames. However, analysis of high-density regions with significant overlaps among the microbubble point spread responses yields high localization errors, constraining the technique to low-concentration conditions. As such, long acquisition times are required to sufficiently cover the vascular bed. In this work, we present a fast and precise method for obtaining super-resolution vascular images from high-density contrast-enhanced ultrasound imaging data. This method, which we term Deep Ultrasound Localization Microscopy (Deep-ULM), exploits modern deep learning strategies and employs a convolutional neural network to perform localization microscopy in dense scenarios. This end-to-end fully convolutional neural network architecture is trained effectively using on-line synthesized data, enabling robust inference in-vivo under a wide variety of imaging conditions. We show that deep learning attains super-resolution with challenging contrast-agent densities, both in-silico as well as in-vivo. Deep-ULM is suitable for real-time applications, resolving about 70 high-resolution patches (128x128 pixels) per second on a standard PC. Exploiting GPU computation, this number increases to 1250 patches per second.