ultrasound-metrics 0.1.0

Metrics for ultrasound data and images

ultrasound-metrics

⚠️ Alpha Release

This library is currently under active development and is released as an alpha version. The primary goal of this release is to collect community feedback.

Introduction

ultrasound-metrics is an open-source Python library for ultrasound data and image quality analysis developed at Forest Neurotech. It is written in the Array API standard, making it compatible across NumPy, JAX, and PyTorch backends. The implementation supports GPU acceleration (CuPy, JAX, PyTorch), software acceleration (JIT or AOT compilation), and interactive region-of-interest selection for metrics in 2D.

Documentation on ultrasound-metrics can be found here and examples can be viewed here. We are actively taking requests for additional metrics that may be helpful to ultrasound researchers.

Installation

Install from PyPi (recommended:)

pip install ultrasound-metrics

Build from source

make install

Build prerequisites:

• uv >= 0.6.10
• optional: make

Documentation

We currently support the following ultrasound data and image quality metrics:

• contrast-to-noise ratio (CNR)
• generalized contrast-to-noise ratio (gCNR)
• signal-to-noise ratio for raw radiofrequency signals (RF SNR)
• temporal signal-to-noise ratio (tSNR)
• sharpness (tenengrad)
• coherence factor and more!

We are actively taking requests for metrics, ultrasound data file types to support, and additional features that would be helpful to the ultrasound imaging community. To make a feature request, please submit a GitHub issue.

Acknowledgements

ultrasound-metrics builds upon the excellent work of the ultrasound imaging community:

• ultraspy - For educational examples and validation benchmarks
• PICMUS - For public, standardized datasets used in examples

This package was developed by the Forest Neurotech team, a Focused Research Organization supported by Convergent Research and generous philanthropic funders.