Chapter 8: Color and Spectral Doppler

Doppler Effect
Doppler Effect
 Doppler Principles

receiver receiver
source

Relative motion of the source causes a

change in received frequency.
Blood flow velocity is measured by
detecting Doppler frequency shifts.
 Doppler Principles

When both source and receiver are

stationary:
When source is stationary and receiver is
moving:
 Doppler Principles

When receiver is stationary and source is

moving:
Doppler Equations
 Doppler Ultrasound
Primary scattering site: red blood cell.
The platelet is too small and the number
of leukocytes is not significant.
The red blood cell size is around several
microns. Thus, scattering and speckle are
also present.
The red blood cells in a sample volume
are assumed to move in unison.
 Doppler Equations

v
c
Typical physiological flows (5-10m/sec at most) are much
slower than sound velocity in the body (~1500m/sec).
Doppler shift is doubled due to round-trip propagation.
Only parallel flows can be detected.
Flow Pattern v. Velocity Profile

ultrasound
beam

0 velocity or
Doppler shift
Flow Pattern v. Velocity Profile

ultrasound
beam

0 velocity or
Doppler shift
Flow Pattern v. Velocity Profile

ultrasound
beam

0 velocity or
Doppler shift
 Doppler Spectrum Estimation
original Doppler shifted demodulated

fs fs+fd fd

Short-time Fourier transform (Spectral

Doppler).
Correlation based estimation (Color Doppler).
 Blood Flow Measurements
Short Time Fourier Analysis: PW,CW,
Audio Doppler.
Correlation processing: Color Doppler,
Doppler Power, Doppler Energy,
Convergent Color Doppler.
Tracking.
Correlation Analysis.
 Continuous Wave (CW) Doppler
T R Received signal

time

time
velocity
 CW Doppler
CW
oscillator spectral
filter estimation
CW audio
transmitter demodulator conversion signal
processing
amplifier

T R speaker display
 CW Doppler

Array CW and AUX CW (half transmit, half

receive).
Mainly for Cardiology.
Good velocity (frequency) resolution.
No range resolution. Flows along the same
direction are all detected.
Frequency downshift due to attenuation can be
ignored.
 CW Doppler Processing

original spectrum
-f0-fd -f0 f0 f0+fd

demodulated
-2f0-fd fd

demodulated and
filtered
fd
 Wall Filter (Clutter Filter)
time
velocity
 CW Doppler Processing
Time-interval histogram.
32-128 pt FFT.
Model-based spectrum estimation (AR),
time-frequency analysis.
Magnitudes are converted in dB and
displayed.
Post-processing similar to B-mode.
 Audio Doppler

For typical blood velocities and carrier frequencies,

the Doppler shifts from blood happen to be in the
human audible range (near DC to 20KHz).
Positive shifts in one channel and negative ones in
the other.
Right
Hilbert transform.
Clinically useful. Left
 CW PW
CW: No range resolution.
Sampling in time = sampling in range.
CW Doppler to PW Doppler.
Pulsed Wave (PW) Doppler

time

PRI
Another View for PW Doppler,
Autocorrelation Processing
www.acuson.com

PRI (Pulse Repetition Interval)

S(t)

Autocorrelation function:
 PW System Diagram
CW
oscillator filter spectral
estimation
sample&hold
PW transmitter audio
(gated CW) demod./LPF conversion signal
processing
gating

transducer amplifier speaker display

Pulsed Wave (PW) Doppler

time

PRI
 Velocity Ambiguity

no aliasing aliasing

vmax -vmax vmax

Velocity Ambiguity (no aliasing)
Velocity Ambiguity (with
aliasing)
 Range Ambiguity

PRI
 Range Ambiguity

PRI T0

OR ?
 Pulse Wave (PW) Doppler

Pulse-echo method, similar to B-mode.

Post-processing similar to CW.
Adjustable range resolution (gate).
Maximum detectable velocity is /(4*PRI).
Maximum depth is (c*PRI)/2.
32-128 point FFT.
PW Blood Flow Measurements
www.acuson.com

PRI (Pulse Repetition Interval)

S(t)
 PW Color Doppler
Single gate multiple gates.
Local flow information 2D flow
information.
Less time for velocity estimation:
quantitative qualitative.
 Color Doppler Parameters

f
Use efficient time domain correlation techniques to
calculate flow characteristics.
Auto-correlation of the Doppler signal.
Commonly derived parameters are mean velocity
(including directionality), variance and energy
(power).
 Color Doppler

multiple gates

multiple firings
 Color Doppler

multiple gates

multiple firings
 Color Doppler

Similar to B-mode, except that each line

is fired multiple times (5-15).
Correlation processing.
Multiple range gates along each line.
Real-time two-dimensional flow imaging.
Poor velocity (frequency) resolution.
Color Doppler Derivation:
Mean Velocity
Color Doppler Derivation:
Mean Velocity
: even : odd
In practice, R(t) is by definition a discrete signal,
Color Doppler Derivation: Mean Velocity
For a single scatterer:

The flow direction is determined by the sign of the

mean frequency.
Color Doppler Derivation: Variance
Color Doppler Derivation: Variance
Color Doppler Derivation: Energy

f
 Color Doppler

Flow parameters are mapped into colors

for display (1D or 2D).
Choice of map affects the presentation of
Color Doppler images.
 Color Doppler: Signal Processing

beam high pass auto- parameter signal and

former filter correlator estimator display processor

Significant frame rate reduction.

Small color boxes are often used to increase
frame rate.
Sophisticated systems utilize multiple beam
formation to further increase frame rate.
 Doppler: Complications
Non-trivial wall filters are required to remove
interference from slow-moving objects.
Adequate signal processing capabilities and
sufficient dynamic range are necessary to
detect weak flows.
Conflicts with frame rate requirements.
Only parallel flow is detectable.
 Homework #3

Due 12:00pm 6/13/2012 by emailing to

[email protected]
[email protected]
 Load hw3_dat.mat. In this data file,
DopplerData represents a PW Doppler data
set. The size of the data is 128 X 400,
representing gated received data with 128
firings and 400 data samples for each firing.
In the same data file, the following
parameters are also included:
PRI: pulse repetition interval in usec.
fs: sampling frequency in MHz.
f0: center frequency of the transmitted pulse in
MHz.
 For each firing, perform the following
processing. The total number of
samples will reduce from 400 to a
single complex sample after the
processing. Repeat the processing for
all firings in order to obtain 128
complex samples.
demodulation low pass filtering summing
1. Based on the 128 complex samples,
calculate the velocity profile using FFT.
The profile should be plotted on a
logarithmic scale with a 40dB range.
The horizontal axis should also be
labeled with proper velocity scale.
(30%)
2. Re-do Problem 1 by using only the even
samples. Also re-do Problem 1, by
using only the first 200 RF data points
during the summation. Explain your
results. (30%)
3. Based on the original complex samples,
calculate and plot the mean velocity, variance
and energy as the number of samples
increases from 2 to 128. Also compare with
the results when only using the center 100
(151:250) points for the summing. Note that
the discrete correlation function is defined as
the following (40%).

4. (bonus) Explore any related Doppler issues

based on the dataset.