博士学位論文要約（平成 23 年 9 月）
Imaging of Transverse Cross Section of Carotid Artery Using Diverging Transmit
Beams from Linear Array Ultrasonic Transducer with Multiple Steered Receive
Beamforming
Akinlolu PONNLE
Supervisor: Hiroshi KANAI, Research Advisor: Hideyuki HASEGAWA
Transverse cross sectional ultrasound imaging of the intima-media complex of the carotid arterial wall is very
difficult using conventional linear scanning. The extent of the arterial wall that is clearly imaged is very limited. The
use of multi-element diverging beam from a linear array transducer with parallel steered receive beamforming and
scanning in a linear sequential fashion is proposed. The B-mode image is reconstructed from combinations of many
steered receiving beams from multiple transmissions per frame. In order to further suppress receive grating lobe
artifacts in re-constructed B-mode images, a technique is proposed, which involves modulating the receiving beams by
a factor that is governed by the envelope of a corresponding signal, formed by filtering the receiving beam by a
bidirectional zero phase low-pass filter with a cut-off frequency that is determined by the receive-steering angle.
Using these methods, offline reconstructed B-mode images from simulations, silicone-rubber tube phantom and in
vivo imaging of the carotid artery showed that the extent of the clearly imaged region is wider than that of conventional
linear scanning. Also, receive-grating lobe artifacts were suppressed without significant loss in spatial resolution.
1. Introduction
In Fig. 1, fd is the distance from the virtual acoustic point
source to the array surface, and L is the transmit
sub-aperture length.
Since the transmitted beams are wide, strong reception
is possible from the directions of the receive-main lobe
and receive-grating lobes leading to grating lobe artifacts.
Therefore, a method of modulating the receiving beams
was proposed to further suppress grating lobe artifacts.
Ultrasound imaging of the intima-media complex of
the carotid arterial wall in the transverse cross-sectional
plane is very difficult using conventional linear scanning
as the extent of the arterial wall that are clearly imaged
is very limited. The intima-media region of the artery is
a useful place to probe in the diagnosis of atherosclerosis
[1].The angular width of the limited imaged region is a
measure of the extent of the anterior and posterior walls
of the arterial cross-section that are clearly imaged, if the
transverse cross-section of the arterial lumen is assumed
to be circular.
The use of multi-element diverging beam from a
linear array transducer with parallel steered receive
beamforming and scanning in a linear sequential fashion
is proposed in order to obtain a wider angular width of
imaged region without steering of transmit beams and at
the same frame rate as in conventional linear scanning.
This type of beam has previously been investigated by
some researchers for diverse purposes [2,3]. B-mode
image was reconstructed from combinations of many
steered receiving beams from multiple transmissions per
frame. The transmission and reception scheme of the
proposed method is illustrated in Fig. 1.
Beam-spread
angle Φ
2.
Principle
The principle of the proposed imaging method using
multi element diverging transmit beams is summarized
in Fig. 2. In this method, within a transmit-receive event,
diverging beams are transmitted from a linear array
transducer and echoes are received. From the echoes
received, steered receiving beams y(t,θ) are created
offline at regular angular intervals (θ is the receive
steering angle). The created, steered receiving beams are
then modulated by modulating weights W(t,θ) before
being spatially combined for image formation.
Transmission
of Diverging
Beams
L
Diverging
transmit
beam
Modulation of
Receiving Beams
by W(t,θ)
Combination of
Multiple
Receiving Beams
V Virtual acoustic
point source
fd
Creation of
Steered Receiving
Beams, y(t,θ)
p
y1(p)
x
W1(p)
Linear transducer
array probe
p
y2(p) x
W2(p)
Σ
Image
formation
at p
Steered
receiving beams
Diverging
wavefronts
p
yNT(p) x WNT(p)
Figure 2 Overview of the principle of the proposed
imaging method.
Figure 1 Illustration of the transmission-reception
scheme of the proposed method.
1
-5
0
5
10
15
0
-100
4
4.5
5
5.5
-100
12
-15
8 -15
-10-10
-15
-10
-10-10
(a)
15-15
12
-15
-15
5 5
10 10
15 15
At-5
Atdepth
depthof0of1520mmmm5
At depth
depth ofof10
15mm
mm
At
0 0
10
15
15 15
1515
50
(a)
0
At depth
depth ofof15
20mm
mm
At
kPa
200
-5
0
5
10
11
12
13
14
15
8
8
8.5
At (x,z) =Trailing
(0,10) mm
pulses
8.5
9
9.5
10
10.5
11
11.5
12
At (x,z) = (0,20) mm
4
4.5
5
5.5
6
6.5
7
7.5
8
8.5
At (x,z) = (0,15) mm
100
11
11.5
12
12.5
13
13.5
14
14.5
15
15.5
Time [s]
-50
-100
Main pulses
7.5
-200
15
-5
00
55
1010
-5 distance
Lateral
[mm]
At depth
of 20 mm[mm]
(b)
Lateral
distance
(1)
-10
7.5
Trailing pulses
8
8.5
9
9.5
10
10.5
11
11.5
12
At (x,z) = (0,20) mm
0
-10
7.5
200
10 10
5 5
7
0
Main pulses
-100
5
At-5-5depthdistance
of0020 mm5[mm]
Lateral
1010
0 0
6.5
100
kPa100
200 1000
0
-200
0
0
-100
-200 -100
-5 -5
-10
-10
kPa
Pressure [kPa]
Time [s]
Time [s]
[s]
[s]
Time
Time
[s]
Time
Time [s]
[s]
Time
11
8 12
1384
10
14
6
15
10
12
-15
8 -15
12
-15
-15
11
12
11
13 8
12
14
13
15 10
14
-5 -5
6
At (x,z) = (0,15) mm
0
[kPa] [kPa]Pressure [kPa]
Pressure [kPa]Pressure
Pressure
[s]
Time [s]
Time
[s]
Time
100
100
8
4
10
6
8
Time [s]
Pressure [kPa]
-10
kPa
6
1515
100
0
-100
(b)
11
11.5
12
12.5
(2)
13
13.5
14
Time [s]
14.5
15
15.5
Figure 4(1) (a) Simulated emitted waves and (b)
AcousticLateral
pressure
as measured with the hydrophone, at
distance [mm]
axial depth of 20 mm. 4(2) The emitted pressure
waveform as it passes a point below the center of the
aperture at an axial depth of 20 mm. (a) Simulation (b)
Measurement (waveform along the white lines in (1-a)
and (1-b)).
-15
-10
-5
0
5
10
15
Acoustic field characteristics of the transmit beam
from linear array transducer was investigated by
simulation using Field II program [6], for beams of
various beam-spread angles, Φ ranging from 0o to 180o.
The simulated emitted acoustic waveform was found to
be a composite of two components which in this
research are referred to as ‘main pulses’ and the ‘trailing
pulses’ (Figs. 4(2-a) and 4(2-b)).
The main pulses are effective for imaging, but the
trailing pulses affect the image contrast and cause depth
sidelobe and range ambiguity artifacts in B-mode images.
The trailing pulses depend on the beam parameters such
as the aperture length, inter-element pitch and the
beam-spread angle. As the beam-spread angle increases,
the absolute amplitude of the main pulses reduces while
those of the trailing pulses increases. The spatial and
temporal length, as well as the amplitude of the main
and trailing pulses, vary for different locations within the
beam, but is prominent within the region defined by the
boundaries of the two edges of the transmit sub aperture
(Fig. 5).
Figure 3 The modulating technique.
The cut-off frequency fc of the ZP-LPF is expressed as
c
,
(1)
fc 
p1  sin  
where c is the speed of ultrasound in the medium, p is
the inter-element pitch, and θ is the receive steering
angle.
The modulating signal RF(n,θ) is obtained by operation
A1 defined by
(2)
otherwise
where yenv(n,θ) is the envelope of the original receiving
beam signal y(n,θ), and yf(env)(n,θ) is the envelope of the
ZP-LP filtered signal yf(n,θ). The output signal ym(n,θ) is
the modulated receiving beam signal given by
(3)
ym (n, )  RF (n, )  y(n, ) ,
Φ = 870
Φ = 1800
0
10
-20
15
-30
20
-40
25
30
-50
-50
-40
-30
-20
-10
0
10
20
30
40
50
(a)
Depth [mm]
0
5
-20
15
-30
20
-40
25
30
-50
-40
-30
-20
-10
0
10
20
30
40
50
Depth [mm]
-20
15
-30
20
-40
25
30
-50
dB
-50
-40
-30
-20
-10
0
10
20
30
40
50
Trailing Pulses field
-50
0
5
dB
-10
10
-20
15
-30
20
-40
25
30
-50
-40
-30
-20
-10
0
10
20
30
40
50
-50
(b)
TP-MP Ratio
(c)
-10
10
dB
-10
10
(b)
3. Computer Simulations
3.1 Acoustic Field
The acoustic field of the diverging beam in water
from a linear array transducer was simulated using Field
II program [6]. Also, the acoustic field in water was
measured using a hydrophone (Toray, H025-002), and
α-10 ALOKA diagnostic ultrasound equipment with a 10
0
5
(a)
Trailing Pulses field
where θ is the receive beam steering angle and (⊗)
connotes multiplication.
Main Pulses field
dB
-10
Depth [mm]
Depth [mm]
Main Pulses field
5
Depth [mm]
y f (env) (n, )  yenv (n, ) ,
At (x,z) = (0,10) mm
8
4
yenv(n,θ)
RF(n,θ)
env.
(modulating signal)
A1
det.
yf(n,θ) yf(env)(n,θ)
ZP-LPF
env.
det.
(fc ∝ θ)
if
6
-15
ym(n,θ)
(modulated signal)
 y f (env) (n, )

RF (n, )   yenv (n, )

1

4
2.5
2
15
1.5
20
1
25
0.5
30
-50
-40
(1)
-30
-20
-10
0
10
TP-MP Ratio
3
5
10
20
30
Lateral distance [mm]
40
50
0
Depth [mm]
y(n,θ)
(original signal)
MHz, 192-element linear array transducer (UST-5545,
Aloka Co., Ltd., Tokyo, Japan). Figure 4(1-a) shows the
simulated Atultrasonic
time-pressure waves and Fig. 4(1-b)
depth of 10 mm
shows the acoustic pressure as measured by the
hydrophone, at a depth of 20 mm below the transducer
surface (active
transmit sub-aperture of 96 elements and
AtAtdepth
depthof of1015mmmm
beam-spread
Φ of 87o). The time axis is referenced
At depthangle
of 10 mm
to the time of the last element(s) to be fired, which are
the two extreme elements in the transmit sub aperture.
Time [s]
In multi-element diverging beam production, the
excitation of each of the elements that make up the
transmit sub-aperture is time-delayed so that spherical
resultant wavefronts are produced in front of the
transducer array over a limited angle. During scanning,
coincident active transmit/receive sub-apertures are
stepped across the array at a pitch of one element.
For one transmit beam from a transmit sub-aperture,
several steered receiving beams in parallel are formed.
The maximum steering angle of receiving beams in each
transmit beam is limited to ±65 degrees. Angular
separation of steered receiving beams is 0.5o. B-mode
image was reconstructed from combinations of many
steered receiving beams at each image point of interest
from multiple transmissions per frame.
Grating lobe artifacts was further suppressed by a
technique of modulating the receiving beams by a factor
that is governed by the envelope of a corresponding
signal, formed by filtering the receiving beam by a
bidirectional zero phase Butterworth low-pass IIR filter
(ZP-LPF) [4,5], with a cut-off frequency that is
determined by the receive-steering angle. The technique
is described in Fig. 3.
3
5
2.5
10
2
15
1.5
20
1
25
30
-50
0.5
-40
-30
(c) (2)
-20
-10
0
10
20
30
40
50
0
Lateral distance [mm]
Figure 5 Simulated acoustic fields (from transmit sub
aperture of 96 elements), separated into fields due to (a)
the main pulses, (b) the trailing pulses and (c) ratio of
2
the maximum amplitude of the main pulses to the
maximum amplitude of the trailing pulses at every point
for (1) Φ of 87o, and (2) Φ of 180o. Similar figures
obtain for sub aperture sizes of 56 and 36 elements.
Maximum observable angular width of 85o with
beam-spread angle of 87o and transmit sub aperture size
of 36 elements was obtained from simulations (Fig.
7(2)).
Simulation also shows that within this region, the
spatial average ratio of the maximum amplitude of the
trailing pulses to the maximum amplitude of the main
pulses is less than one (1) for beam-spread angles up to
120o (Fig.6).
Average TP-MP Ratio between 10 mm and 20 mm of Region R
4. Phantom Experiments
A silicone-rubber tube phantom in water was scanned
in transverse cross sectional plane using conventional
linear scanning and the proposed method for beams with
beam-spread angles ranging from 87o to 150o; and sub
apertures sizes of 36, 56 and 96 elements. The scanning
was performed using α-10 ALOKA ultrasound
equipment and a 10 MHz, 192-element linear array
transducer probe (UST-5545, Aloka Co., Ltd., Tokyo,
Japan). Received ultrasonic RF echoes were sampled at
a frequency of 40 MHz. Number of transmissions for
each sub-aperture size is given by equation (4).
Figure 8(1) shows the reconstructed B-mode images
of the silicone-rubber tube phantom. The plot in Fig.
8(2) shows the variation of the angular width (AW) of
the clearly imaged region with different beam-spread
angles, Φ, for the silicon-rubber tube phantom in
comparison with the angular width obtained with
conventional scanning. Maximum angular width of 60o
with beam-spread angle of 87 o and transmit sub aperture
size of 36 elements was obtained from diverging beam
scanning as opposed to 24o obtained from conventional
linear scanning. This is an increase of about 2.5.
Value
pulses
to Sub
main
pulses of the acoustic beam between
0.8
aperture size: 36 elements
Submm
aperture
elements
depth
andsize:
2056mm
below the transducer surface
0.6 of 10
Sub aperture size: 96 elements
for0.4sub aperture sizes of 36, 56 and 96 elements.
0.2
3.2 0Virtual Scanning
20
40
60
80
100
140
160
A 0 2-interface
cylindrical
tube120composed
of180point
 [deg]
scatterers of inter-scatterer distance of 20 μm (outer wall
diameter of 10 mm and inner wall diameter of 8 mm)
was simulated, and virtual scanning in the transverse
plane, using conventional linear scanning and the
proposed scheme with a 192-element linear array
transducer, were performed for various beam-spread
angles and sub apertures sizes of 36, 56 and 96 elements
using Field II program [5]. The center of the tube’s cross
section is positioned at a depth of 14.1 mm.
Number of transmissions, NT, for each sub-aperture
size is given by
NT  N ea  N es  1 ,
(4)
10
10
12
12
14
16
18
18
20
20
-5
0
5
0
5
Lateral distance [mm] Lateral distance [mm]
(1) (a)
(1)
70
14
16
18
(b3)
(b2)
60
Sub aperture size: 36 elements
Sub aperture size: 56 elements
Sub aperture size: 96 elements
Conventional (96 elements)
20
40
60
80
100
120
140
12
4
6
8
10
Depth [mm]
16
18
20
42
64
68
8
10
14
16
16
18
18
20
20
20
22
22
22
20
16
16
18
18
20
22
10 0
20
0
14
14
20
42
160
64
86
108
10 0
14
16
18
20
22
42
64
86
108
0
10
distance
[mm][mm]
Lateral
distance
[mm][mm]Lateral
Lateral
distance
distance
[mm][mm]Lateral
Lateral
distance
Lateral
distance
(a)
(b2)
(b1)
2
(b3)
Sub aperture
Sub aperture
Sub aperture
Conventional
size: 36 elements
size: 56 elements
size: 96 elements
(96 elements)
40
30
100
120
140
160
1
1
0.9
0.9
0.8
0.8
0.7
5.
0.7
0.6
100
120
140
160
0.5
80
100
120
140
In vivo Imaging
A beam-spread angle of 90o with transmit sub
aperture size of 36 elements were selected for in vivo
imaging. In vivo imaging of the transverse cross section
of the carotid artery of three human subjects (adult
males) were taken using the α-10 ALOKA ultrasound
equipment and the 192-element linear transducer array
probe used in the phantom experiments. Two scans
were performed in each experiment; conventional
180
Beam-spread angle [deg]
(2)
Figure 7(1) B-mode images of the simulated cylindrical
tube. (a) Conventional scanning, (b) Diverging beam
scanning, beam-spread angle Φ of 87o and sub aperture
size of (b1) 96, (b2) 56, and (b3) 36 elements. 7(2)
Variation of angular width (AW) of the imaged region
for different Φ and sub aperture sizes of 36, 56 and 96
elements.
3
160
4
6
Lateral distance
Angular Width (phantom)
50
0.5
80
0
10
12
14
0.6
20
10
12
(2)
CR [same
receive sub aperture
of the
silicone-rubber
tube (96)]
phantom. (a) Conventional scanning. (b) Diverging
beam scanning (b1) Φ of 87o and sub aperture of 96
elements, (b2) Φ of 90o and sub aperture of 56 elements,
(b3) Φ of 90o and sub aperture of 36 elements. 8(2)
Variation of angular width (AW) of the clearly imaged
regionBeam-spread
for different
and coincident
transmit/receive
angleΦ,
[deg]
Beam-spread
angle [deg]
sub-aperture sizes of 36, 56 and 96 elements.
5
80
40
10
12
Beam-spread angle [deg]
CR B-mode
(phantom) images
Figure 8(1)
Lateral distance [mm]
(b1)
18
14
60
20
80
0
Depth [mm]
Depth [mm]
Depth [mm]
2
Lateral distance [mm]
12
-5
16
20
0
Angular width vs beam-spread angle from simulation
0
18
12
14
22
20
-5
16
CR
16
Depth [mm]
14
AW
10
12
CR
10
14
Angular Width [deg]
10
Depth [mm]
8
Angular Width [deg]
Depth [mm]
8
12
20
Conventional
10
10
12
where Nea is the number of elements in the array and Nes
is the number of elements in the sub aperture.
8
8
8
10
Depth [mm]
Sub aperture size: 36 elements
Sub aperture size: 56 elements
Sub aperture size: 96 elements
0
0
20
40
60
80
100
120
140
160
180
 [deg]
Average
TP-MP Ratio
between
10 mmamplitude
and 20 mm ofof
Region
Figure
6 Average
ratio
of the
theRRtrailing
1
0.2
Depth [mm]
0.4
Depth [mm]
Value
0.6
Depth [mm]
1
0.8
focused linear scanning and the proposed method
(diverging beam scanning). Received RF signals were
sampled at a frequency of 40 MHz and the elements’
received RF data for each transmission were acquired for
offline processing.
Figures 9(a) and 9(b) shows the re-constructed
B-mode images obtained from the carotid artery
scanning of the first subject. Figure 9(b) was
re-constructed from 141 sequential transmissions
(symmetrical about center of lumen) per frame. Wider
regions of the posterior intima and adventitia can be seen
in the diverging beam image than in the conventional
focused beam image. The average result obtained (from
the three subjects) for the angular width from diverging
beam scanning is about 45o and 24o from conventional
linear scanning; yielding a ratio of about 1.9.
16 16
17
24o
19
20 20
21 21
22
2
4
6
8
18
19
20
2 2
4 4
6 6
8 8
2
Lateral
distance
[mm]
Lateral
distance
[mm]
4
Magnitude
Depth [mm]
10
Depth [mm]
20
(a) 0
10
Magnitude [dB]
10
20
15
20
(b) 0
10
20
10
15
20
Lateral Distance [mm]
30
25
30
25
30
-50
-100
0
5
10
15
20
Lateral Distance [mm]
(2)
0
-50
-100
(a)
rcv beams
ZP-LPF beams
mod-rcv beams
10
1
12
14
16
18
Depth [mm]
Positions of the scatterers
20
20
0
10
20
30
Lateral distance [mm]
(1)
(b)
10
12
14
16
Depth [mm]
(c)
8
18
Depth [mm]
10
20
Lateral distance [mm]
(1)
10
20
30
Lateral distance [mm]
ZP-LPF Receiving Beams
10
15
20
(b)
0
10
20
Lateral distanceBeams
[mm]
Mod-Receiving
30
0
10
20
Lateral distance [mm]
30
10
15
20
30
(c)
(2)
References
1) H. Hasegawa and H. Kanai, IEEE Trans. UFFC, 55
(2008) 1921-1934.
2) M. Karaman, P. Li, and M. O’Donnell, IEEE Trans.
UFFC, 42 (1995) 429-442.
3) G. R. Lockwood, J. R. Talman, and S. S. Brunke,
IEEE Trans. UFFC, 45 (1998) 980-988.
4) F. Gustafsson, IEEE Trans. Signal Process, 44 (1996)
988-992.
5) I. W. Selesnick and C. S. Burrus, Proc IEEE Inter
Conf Acoust Speech Signal Proces, 3 (1996)
1688-1694.
6) J. A. Jensen, Med. Biol. Eng. Comp., 34, (1996)
351-353.
15
(c)
20
0
Conclusion
This research has demonstrated through simulations,
phantom and in vivo imaging that by employing
multi-element diverging beams from a linear array
transducer, improved image of the carotid artery in the
transverse cross-sectional plane can be obtained than is
possible with conventional linear scanning. In this
method, there is no need for steering of transmit beams.
Also, the proposed method of modulation of
receiving beams further suppresses receive-grating lobe
artifacts. It is effective for keeping the phases of the
steered receiving beams unaffected as well for
preserving the spatial resolution of the reconstructed
image.
rcv beams
ZP-LPF beams
mod-rcv beams
30
Magnitude [dB]
Depth [mm]
15
15
20
7.
rcv beams
ZP-LPF beams
mod-rcv beams
5
30
15
Figure 11 Reconstructed B-mode images of (1)
silicone-rubber tube phantom in the transverse
cross-sectional plane, and (2) carotid artery of the first
subject in the transverse cross sectional plane (a) from
receiving beams, (b) from ZP-LPF signals and (c) from
modulated receiving beams. All images are in 60 dB
range.
Suppression of Grating Lobe Artifacts
The proposed method of suppression of grating lobe
artifacts by modulation of receiving beams was applied
to reconstructed B-mode image
obtained from virtual
1
scanning of simulated columns of point scatterers in
0.5
water (Fig. 10(1)), and a grating
lobe artifact suppression
of about 40 dB was realized (Fig.
10(2)).
0
0
6
10
20
Lateral distanceBeams
[mm]
Mod-Receiving
10
Lateral distance [mm]
6.
10
0
0
0
(a)
(b)
Figure 9 B-mode images of the carotid artery of the first
subject in the transverse cross sectional plane. (a)
Conventional scanning, (b) Diverging beam scanning, Φ
of 90o and transmit/receive sub-aperture size of 36/96
elements. Images are in 50 dB range.
0
20
10
(a)
Depth [mm]
Depth [mm]
Depth [mm]
17
10
20
30
Lateral distance [mm]
ZP-LPF Receiving Beams
15
(b)
22
0 0
Lateral distance [mm]
0
10
21
22 22
0
Diverge
16
45o
19 19
21
20
15
18 18
20
15
(a)
14
17 17
18
10
Depth [mm]
15 15
16
Receiving Beams
Receiving Beams
Depth [mm]
15
Depth [mm]
Depth [mm]
Depth [mm]
Conv
Diverge
14 14
Depth [mm]
Conv
14
The proposed method of modulation of receiving
beams was also applied to the phantom B-mode image,
and the in vivo images. It suppresses grating lobe
artifacts without loss in spatial resolution (Figs. 11(1)
and 11(2)).
20
(3)
Figure 10(1) Reconstructed B-mode image of virtual
scanning of simulated columns of point scatterers using
diverging beam scanning, (a) from receiving beams, (b)
from ZP-LPF signals and (c) from modulated receiving
beams. 10(2) Lateral profile at a depth of 14 mm.
10(3-a) Depth profile along the scatterers of the central
column, 10(3-b) Positions of the point scatterers of
10(3-a).
4