Personal tools

Minimum Variance Beamforming for Wearable Ultrasound Probes

From iis-projects

Jump to: navigation, search

Short Description

Wearable Ultrasound probes appear as a promising tool for assisting medical therapies and sport science in the characterization of musculoskeletal features on humans. A typical wearable probe is characterized by a reduced nummber of channels: the availability of few channels severely reduces the quality of the acquired ultrasound images. Standard beamforming techniques are thus not well suited to obtain higher image qualities. Minimum Variance Beamforming (MVBF) is an adaptive beamforming technique for sensor arrays. It reduces the noise contribution by minimizing the variance of the output noise. Despite being computationally expensive, it achieves better ultrasound image quality than conventional Delay-and-Sum (DAS) beamformers. The key component of a MVBF algorithm is the covariance matrix of the sensor's signals. While the covariance matrix can be reliably estimated on systems with a large number of channels, its application to a reduced number of channels is still unexplored. In this project, you will explore alternative methods for the estimation of the covariance matrix on ultrasound data.

Goal & Tasks

The goal of this project is to evaluate the performance of Minimum Variance Beamformer and of different techniques for the estimation of the covariance matrix, applied to ultrasound data with a reduced number of channels. The main tasks of this project are:

  • Literature study of ultrasound beamforming (standard Delay-and-Sum beamforming, and minimum covariance beamforming)
  • Literature review of MVBF modifications for operation on a reduced number of channels
  • Perform numerical simulations (Field II) to obtain synthetic ultrasound data
  • Algorithm implementation and testing on the simulated data
  • Application on real ultrasound data


  • Linear algebra
  • Signal processing basics
  • Python (numpy, scipy, etc.)

Status: Available

Looking for Interested Students
Supervision: Sergei Vostrikov, Andrea Cossettini


20% Literature Study
20% Simulations of the Imaging Process
40% Programming and Testing
20% Application on real data

NOTE: in view of COVID-19 restrictions, this project can be done 100% remotely


Luca Benini

↑ top

Practical Details