Baseline: Lasso Implementation for ULM (Ultrasound Localization Microscopy)

Ultrasound Localization Microscopy has enabled super-resolution vascular imaging through precise localization of individual ultrasound contrast agents (micro bubbles) across numerous imaging frames.

However, recovery of dense micro bubble centers is not an easy task due to significant overlaps among the micro bubble point spread function.

Subsequently, long acquisition times are required to sufficiently cover the vascular bed since low density micro bubble solution is used.

Lasso Implementation is chosen as the baseline for this problem.

\[ \min_{x} || Ax – y ||_{2}^{2} + \lambda || x ||_{1} \]

In order to implement LASSO, PSF function of a micro bubble from the center of the image is considered.


Lasso implementation is tested on Field 2 simulation results. Simulation result is divided into three regions: Near to the Sensor Array, Center of the target scene and Far

Nearest region
Lasso Result
Lasso result
Farthest region
Lasso result

LASSO is also implemented using patches. For simplicity, I just used three patches: near, center and farthest.

PSF for Near

PSF for Center
PSF for Far

Lasso results for the same simulation above are as follows:




Leave a Comment

Your email address will not be published. Required fields are marked *