Biomedical Engineering Reference
In-Depth Information
2. Create a vector h :
1for i
=
1
, (
n
/
2
)+
1
h
(
i
)=
2for i
=
2
,
3
,..., (
n
/
2
)
(2.65)
0for i
=(
n
/
2
)+
2
,...,
n
3. Calculate the element-wise product of X and h : Y
=
X
·
h
4. Calculate the inverse FFT of Y .
This algorithm is implemented in MATLAB Signal Processing Toolbox as
hilbert . The analytic signal can be presented in the form:
x a
(
t
)=
A
(
t
)
cos
(
φ
(
t
))
(2.66)
where A
(
t
)= |
x a
(
t
) |
is called the instantaneous amplitude, φ
(
t
)=
arg
(
x a
(
t
))
and
))
dt is called the instantaneous frequency 1 .
The concepts of instantaneous amplitude and frequency are useful in case of sim-
ple signals without overlapping frequency components, e.g., in Figure 2.12 a-c; the
instantaneous amplitude recovers the modulation of the chirp, and instantaneous fre-
quency recovers its linearly changing frequency. However, the instantaneous mea-
sures turn out to be meaningless, if the signal is more complicated, e.g., if there are
two structures with different frequencies present at the same moment. Figure 2.12 d-f
illustrates such a case. The signal in panel d) is composed of two signals analogous
to that in panel a), but shifted in frequency by a constant. Its instantaneous amplitude
(panel e) and frequency (panel f) display fast oscillatory behavior.
d arg
(
x a (
t
1
f
(
t
)=
2.4.2
Analytic tools in the time-frequency domain
2.4.2.1
Time-frequency energy distributions
Signal energy can be computed in time or in frequency domain as:
Z
Z
2 dt
2 df
E x
=
|
x
(
t
) |
=
|
X
(
f
) |
(2.67)
2 are interpreted as energy density. The idea of sig-
nal energy density can be extended to the time-frequency space. The time-frequency
energy density ρ x (
2 or
and the quantities
|
x
(
t
) |
|
X
(
f
) |
should represent the amount of signal energy assigned to a
given time-frequency point
t
,
f
)
(
t
,
f
)
,andfulfill the conditions:
1. The total energy is conserved:
Z
Z
E x
=
ρ x
(
t
,
f
)
dtd f
(2.68)
1 Note, that in case of a stationary signal, e.g., x = B cos ( ft ) these definitions retrieve amplitude A ( t )= B
and frequency φ
(
t
)=
f of the signal.
 
Search WWH ::




Custom Search