TY - JOUR

T1 - Convex Array Vector Velocity Imaging Using Transverse Oscillation and Its Optimization

AU - Jensen, Jørgen Arendt

AU - Brandt, Andreas Hjelm

AU - Bachmann Nielsen, Michael

PY - 2015

Y1 - 2015

N2 - A method for obtaining vector flow images using
the transverse oscillation (TO) approach on a convex array is
presented. The paper presents optimization schemes for TO
fields and evaluates their performance using simulations and
measurements with an experimental scanner. A 3-MHz 192-element
convex array probe (pitch 0.33 mm) is used in both
simulations and measurements. A parabolic velocity profile is
simulated at a beam-to-flow angle of 90°. The optimization
routine changes the lateral oscillation period λx as a function
of depth to yield the best possible estimates based on the energy
ratio between positive and negative spatial frequencies in
the ultrasound field. The energy ratio is reduced from −17.1
dB to −22.1 dB. Parabolic profiles are estimated on simulated
data using 16 emissions. The optimization gives a reduction in
standard deviation from 8.81% to 7.4% for 16 emissions, with
a reduction in lateral velocity bias from −15.93% to 0.78% at
90° (transverse flow) at a depth of 40 mm. Measurements have
been performed using the experimental ultrasound scanner
and a convex array transducer. A bias of −0.93% was obtained
at 87° for a parabolic velocity profile along with a standard
deviation of 6.37%. The livers of two healthy volunteers were
scanned using the experimental setup. The in vivo images
demonstrate that the method yields realistic estimates with a
consistent angle and mean velocity across three heart cycles.

AB - A method for obtaining vector flow images using
the transverse oscillation (TO) approach on a convex array is
presented. The paper presents optimization schemes for TO
fields and evaluates their performance using simulations and
measurements with an experimental scanner. A 3-MHz 192-element
convex array probe (pitch 0.33 mm) is used in both
simulations and measurements. A parabolic velocity profile is
simulated at a beam-to-flow angle of 90°. The optimization
routine changes the lateral oscillation period λx as a function
of depth to yield the best possible estimates based on the energy
ratio between positive and negative spatial frequencies in
the ultrasound field. The energy ratio is reduced from −17.1
dB to −22.1 dB. Parabolic profiles are estimated on simulated
data using 16 emissions. The optimization gives a reduction in
standard deviation from 8.81% to 7.4% for 16 emissions, with
a reduction in lateral velocity bias from −15.93% to 0.78% at
90° (transverse flow) at a depth of 40 mm. Measurements have
been performed using the experimental ultrasound scanner
and a convex array transducer. A bias of −0.93% was obtained
at 87° for a parabolic velocity profile along with a standard
deviation of 6.37%. The livers of two healthy volunteers were
scanned using the experimental setup. The in vivo images
demonstrate that the method yields realistic estimates with a
consistent angle and mean velocity across three heart cycles.

U2 - 10.1109/TUFFC.2015.006970

DO - 10.1109/TUFFC.2015.006970

M3 - Journal article

C2 - 26670846

SN - 0885-3010

VL - 62

SP - 2043

EP - 2053

JO - IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control

JF - IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control

IS - 12

ER -