Related Experiment Video
Updated: Dec 31, 2025
07:26
Two-Dimensional Super-Resolution Visualization of Rat Brain Microvasculature Using Ultrasound Localization Microscopy
Published on: March 28, 2025
964
Deep Learning-Based Super-resolution Ultrasound Speckle Tracking Velocimetry
Jun Hong Park1, Woorak Choi1, Gun Young Yoon1
1Department of Mechanical Engineering, Pohang University of Science and Technology (POSTECH), Nam-gu, Pohang, Republic of Korea.
Ultrasound in Medicine & Biology
|January 10, 2020
Summary
This study introduces deep learning-based super-resolution ultrasound (DL-SRU) for high-resolution imaging without contrast agents. DL-SRU accurately maps vessel structure and blood flow, offering a robust alternative for clinical ultrasound applications.
Area of Science:
- Medical Imaging
- Biomedical Engineering
- Artificial Intelligence
Background:
- Deep ultrasound localization microscopy (deep-ULM) offers sub-wavelength resolution but relies on contrast agents, posing potential risks.
- There is a need for advanced ultrasound techniques that provide high resolution without the risks associated with contrast agents.
Purpose of the Study:
- To develop and validate a deep learning-based super-resolution ultrasound (DL-SRU) technique for high-resolution imaging of vasculature.
- To demonstrate the capability of DL-SRU to reconstruct vessel geometry and measure blood flow dynamics without the use of contrast agents.
- To compare the performance of DL-SRU with existing ultrasound localization microscopy methods.
Main Methods:
- A convolutional neural network was trained using synthetic tracer images for red blood cell (RBC) localization and vessel reconstruction.
- The DL-SRU algorithm was validated using in silico and in vitro models, comparing vascular profile widths with standard ultrasound localization microscopy (ULM) and average intensity methods.
- A two-frame particle tracking velocimetry (PTV) algorithm was integrated with DL-SRU for accurate blood flow velocity measurements.
Main Results:
- DL-SRU successfully reconstructed high-resolution vessel geometry and mapped RBC positions without contrast agents in ultrasound images.
- Validation showed DL-SRU performance comparable to deep-ULM in localization robustness, computational speed, and measurement accuracy.
- The integrated PTV algorithm accurately measured flow velocity, demonstrating applicability for in vivo conditions.
Conclusions:
- DL-SRU provides a robust, contrast-agent-free method for high-resolution ultrasound imaging of vasculature.
- The technique enables simultaneous mapping of vessel morphology and blood flow dynamics with high accuracy and speed.
- DL-SRU shows significant potential as a valuable tool for clinical practice.
