Getting Started

Getting Started#

Installation#

Install from PyPI (recommended):

pip install ultrasound-metrics

For development installation:

git clone https://github.com/Forest-Neurotech/ultrasound-metrics.git
cd ultrasound-metrics
make install

Quick Start#

Basic usage with NumPy arrays:

import numpy as np
import ultrasound_metrics as um

# Generate example beamformed channel data
n_elements, height, width = 64, 100, 150
channel_data = np.random.complex128((n_elements, height, width))

# Compute coherence factor
coherence = um.coherence_factor(channel_data)
print(f"Coherence shape: {coherence.shape}")  # (100, 150)
print(f"Coherence range: [{coherence.min():.3f}, {coherence.max():.3f}]")

# Compute image sharpness
bmode_image = np.abs(channel_data.sum(axis=0))
sharpness = um.tenengrad(bmode_image)
print(f"Tenengrad sharpness: {sharpness:.2f}")

Multi-Backend Usage#

The same API works with different array libraries:

import jax.numpy as jnp
import torch
import ultrasound_metrics as um

# JAX example with JIT compilation
data_jax = jnp.ones((64, 100, 100), dtype=jnp.complex64)

compute_coherence_jit = jax.jit(um.coherence_factor)
coherence_jax = compute_coherence_jit(data_jax)

# PyTorch example
data_torch = torch.randn(64, 100, 100, dtype=torch.complex64)
coherence_torch = um.coherence_factor(data_torch)

Next Steps#

  • Explore the Concepts to learn about ultrasound metrics

  • Browse the Examples for practical workflows and inspiration

  • Check the API Reference for complete function documentation

  • See Contributing if you’d like to contribute to the project