Media and phantoms

Materials

A Material holds the acoustic and thermal properties of one tissue or phantom material:

Field	Meaning	Unit
c	sound speed	m/s
rho	density	kg/m³
alpha0, y	attenuation \(\alpha_0 f^y\)	dB/(cm MHz^y)
ba	nonlinearity B/A
k	thermal conductivity	W/(m K)
cp	specific heat	J/(kg K)
perfusion	blood mass flow per volume	kg/(m³ s)
echo	backscatter level relative to a reference	dB
source	where the values come from

import numpy as np
import sonosk as sk

gel = sk.Material("gel", c=1540, rho=1000, alpha0=0.5, k=0.55, cp=4000)
gel.impedance            # rho c
gel.attenuation(5e6)     # Np/m at 5 MHz
soft = gel.with_(alpha0=0.7, name="soft")

The library does not yet ship tissue values. A database with a published source for every number is planned for version 0.2 (see the Roadmap); until then give your own values and fill in source.

Media

A Medium turns a label map into property maps on a grid:

from sonosk import phantom

g = sk.Grid.centered((201, 301), 0.1e-3, depth_start=0.0)
labels = phantom.layers(g, [2e-3, 8e-3])    # 0 above, 1, 2, then 3 below
mats = [water, skin, fat, gland]            # one Material per label
med = sk.Medium(g, labels, mats)

med.c, med.rho, med.alpha0                   # arrays with the grid shape
med.mask("fat")                              # where the material is fat
med.volume("gland")                          # area in 2-D, volume in 3-D

Shapes

Every shape returns a boolean mask on the grid. Positions are given as (x, y, z); on a 2-D grid y is ignored.

phantom.sphere(g, center=(0, 0, 20e-3), radius=3e-3)
phantom.ellipsoid(g, center=(0, 0, 25e-3), semi_axes=(5e-3, 5e-3, 2e-3), angle=0.3)
phantom.cylinder(g, point=(0, 0, 15e-3), direction=(1, 0, 0), radius=1e-3)
phantom.box(g, low=(-5e-3, -5e-3, 10e-3), high=(5e-3, 5e-3, 12e-3))
phantom.halfspace(g, depth=30e-3)

paint writes a label into a mask, so phantoms are built from the outside in:

labels = np.zeros(g.shape, int)
labels = phantom.paint(labels, phantom.halfspace(g, 2e-3), 1)
labels = phantom.paint(labels, phantom.sphere(g, (0, 0, 20e-3), 4e-3), 2)

Scatterers for imaging

pos, amp = phantom.scatterers(med, density=60, rng=0)

density is the number of scatterers per mm² on a 2-D grid and per mm³ on a 3-D grid. Ten or more per resolution cell give fully developed speckle. The amplitude of each scatterer is a normal random number times \(10^{\mathrm{echo}/20}\) of its material. point_targets adds wires:

wires, wamp = phantom.point_targets([(0, 0, 30e-3), (5e-3, 0, 30e-3)], amplitude=20.0)

Displaying a phantom

import sonosk.plot as splt
x, z = g.axes()
splt.labels_map(med.labels, x, z, names=[m.name for m in med.materials])