RMSCAT Remove insignificant scatterers for faster simulations

RMSCAT removes the scatterer that would have little effect on an ultrasound image simulated by SIMUS or SIMUS3.

Contents

Syntax

[X2,Y2,Z2,RC2] = RMSCAT(X,Y,Z,RC,DELAYS,PARAM) removes the scatterers that would have little effect on an ultrasound image simulated by SIMUS or SIMUS3 and of dynamic range 40 dB. The retained scatterers are located at {|X2|,|Y2|,|Z2|} and their reflection coefficients are RC2.

RMSCAT can be used before SIMUS or SIMUS3. After using RMSCAT, you can thus consider running RF = SIMUS(X2,Y2,Z2,RC2,DELAYS,PARAM) instead of RF = SIMUS(X,Y,Z,RC,DELAYS,PARAM). Same for SIMUS3.

[X2,Z2,RC2] = RMSCAT(X,Z,RC,DELAYS,PARAM) assumes that you will use 2-D acoustics in the X-|Z| plane: RF = SIMUS(X2,Z2,RC2,DELAYS,PARAM) or RF = SIMUS(X2,[],Z2,RC2,DELAYS,PARAM).

I = RMSCAT(...) returns the linear indices corresponding to the retained scatterers, i.e. X2 = X(I), Y2 = Y(I),...

[...] = RMSCAT(...,DR) removes the scatterers that would have little effect on a simulated ultrasound of dynamic range DR dB. By default, DR = 40.

Example: Ultrasound image of a carotid

This example shows how you can make simulations faster with RMSCAT

Read a carotid image

I = rgb2gray(imread('carotid.jpg'));

Choose a linear array

param = getparam('L11-5v');

Generate scatterers with GENSCAT

[xs,~,zs,RC] = genscat([5e-2 NaN],param,I,.4);

Simulate an ultrasound image with SIMUS (plane wave)

dels = zeros(1,param.Nelements); % all delays are zero (plane wave)
opt.WaitBar = false;
tic
RF1 = simus(xs,zs,RC,dels,param,opt);
t1 = toc; % computation time for SIMUS

Remove some scatterers with RMSCAT

tic
[xs2,zs2,RC2] = rmscat(xs,zs,RC,dels,param);
t2 = toc; % computation time for RMSCAT + SIMUS

Display the scatterers

subplot(211)
scatter(xs*1e2,zs*1e2,2,20*log10(RC/max(RC(:))),'filled')
ylabel('[cm]')
clim([-40 0]) % range = [-40 0] dB
axis ij tight equal
title([int2str(numel(RC)) ' scatterers'])
set(gca,'XColor','none','box','off')
subplot(212)
scatter(xs2*1e2,zs2*1e2,2,20*log10(RC2/max(RC2(:))),'filled')
ylabel('[cm]')
clim([-40 0]) % range = [-40 0] dB
axis ij tight equal
title({'after RMSCAT:',[int2str(numel(RC2)) ' scatterers']})
set(gca,'XColor','none','box','off')
colormap(hot)

Simulate again for a comparison

tic
[RF2,param] = simus(xs2,zs2,RC2,dels,param,opt);
t2 = t2 + toc;

Demodulate with RF2IQ

IQ1 = rf2iq(RF1,param.fs,param.fc);
IQ2 = rf2iq(RF2,param.fs,param.fc);

Beamform with DAS

[x,z] = meshgrid(linspace(-2e-2,2e-2,256),linspace(eps,3e-2,200));
param.fnumber = [];
IQ1b = das(IQ1,x,z,dels,param);
IQ2b = das(IQ2,x,z,dels,param);

Create B-mode images with BMODE

I1 = bmode(IQ1b,50); % dynamic range = 50 dB
I2 = bmode(IQ2b,50);

Display the images, and compare

subplot(211)
imagesc(x(1,:)*100,z(:,1)*100,I1)
ylabel('[cm]')
axis equal ij tight
title(['computation time: ' int2str(t1) ' s'])
subplot(212)
imagesc(x(1,:)*100,z(:,1)*100,I2)
ylabel('[cm]')
axis equal ij tight
title({'(with RMSCAT)',['computation time: ' int2str(t2) ' s']})
colormap gray

See also

genscat, simus, simus3

About the author

Damien Garcia, Eng., Ph.D.
INSERM research director
Creatis, University of Lyon, France

websites: BioméCardio, MUST

Date modified

Published with MATLAB® R2024a