Medical image processing systems and methods thereof
The present invention provides improved medical imaging apparatuses, systems and methods, especially those using ultrasonography. For example, an imaging system can include a transducer, a scanner, a signal processing unit, and a display unit. The scanner, coupled to the transducer, can convert analog signals to a plurality of digital signals. The signal processing unit, coupled to the scanner, can be configured to: receive the plurality of digital signals; multiply a first magnitude estimate of the plurality of digital signals by a probability that an actual magnitude of the plurality of digital signals is greater than zero to determine a second magnitude; and form an image based on the second magnitude estimate. The display unit displays the image. The present invention also provides for products produced by the methods of the present invention and for apparatuses used to perform the methods of the present invention.
1. An ultrasound imaging system comprising:
a transducer;
a scanner, operatively connected to the transducer, for converting a plurality of analog signals from the transducer to a plurality of digital signals;
a signal processing unit, operatively connected to the scanner, the signal processing unit configured to:
receive the plurality of digital signals, the plurality of digital signals being correlated to a magnitude of a pixel;
determine a first magnitude estimate of the plurality of digital signals, the first magnitude estimate being indicative of a standard magnitude estimate;
determine a second magnitude estimate by subtracting a standard deviation of noise from the first magnitude estimate;
determine a third magnitude estimate by multiplying the first magnitude estimate and a probability that an actual magnitude of the plurality of digital signals is greater than 0;
determine a fourth magnitude estimate by averaging the second magnitude estimate and the third magnitude estimate
compare error rates for true magnitude values in determining the second, third, and fourth magnitude estimates to select a reliable magnitude estimate; and
form at least one portion of an image based upon the reliable magnitude estimate thereby providing an enhanced contrast resolution of the image compared to an image created by the first magnitude estimate and a display unit, connected to the signal processing unit, showing the image;
wherein the first magnitude estimate of the plurality of digital signals is generated using the formula:
v
^
=
1
N
∑
i
v
i
wherein
{circumflex over (ν)} is the first magnitude estimate,
N is a number of samples, and
ν i is the i th input magnitude;
wherein the second magnitude estimate is generated using a formula:
v
^
=
Real
{
1
N
∑
i
v
i
2
-
2
σ
2
}
wherein,
{circumflex over (ν)} is the second magnitude estimate,
N is a number of samples,
ν i is an i th input magnitude, and
σ is a standard deviation of noise in Gaussian distribution;
wherein the third magnitude estimate is generated using either one of a formula:
v
^
=
(
1
N
∑
i
v
i
)
(
1
-
e
1
N
∑
i
-
v
i
2
/
2
σ
2
A
)
or a quadrature specific formula:
v
^
=
(
Q
2
+
I
2
)
(
1
-
e
∑
i
-
(
Q
i
2
+
I
i
2
)
/
2
σ
2
A
)
wherein,
{circumflex over (ν)} is the third magnitude estimate,
N is a number of samples,
ν i is an i th input magnitude, and
σ is a standard deviation of noise in Gaussian distribution;
A is a constant for a filter aggressiveness,
Q is quadrature components, and
I is in phase components; and
wherein the probability is reduced to:
p
(
v
>
0
|
Q
,
I
)
=
1
-
∏
i
e
-
Q
i
2
/
2
σ
2
A
∏
i
e
-
I
i
2
/
2
σ
2
A
wherein
σ is a standard deviation of noise in Gaussian distribution, and
A is a constant for a filter aggressiveness.
2. The imaging system of claim 1 , wherein the demodulation is a time-domain demodulation formulation, a frequency domain demodulation formulation, or a combination thereof.
3. A method for forming an ultrasound image comprising:
receiving, at one or more processors, a plurality of digital signals, the plurality of digital signals being correlated to a magnitude of a pixel;
determining, via the one or more processors, a first magnitude estimate of the plurality of digital signals, the first magnitude estimate being indicative of a standard magnitude estimate;
determining, via the one or more processors, a second magnitude estimate by subtracting a standard deviation of noise from the first magnitude estimate;
determining, via the one or more processors, a third magnitude estimate by multiplying the first magnitude estimate and a probability that an actual magnitude of the plurality of digital signals is greater than 0;
determining, via the one or more processors, a fourth magnitude estimate by averaging the second magnitude estimate and the third magnitude estimate;
comparing error rates for true magnitude values in determining the second, third, and fourth magnitude estimates to select a reliable magnitude estimate;
forming at least one portion of an image based upon the reliable magnitude estimate thereby providing an enhanced contrast resolution of the image compared to an image created by the first magnitude estimate; and
wherein the first magnitude estimate of the plurality of digital signals is generated using a formula:
v
^
=
1
N
∑
i
v
i
wherein,
{circumflex over (ν)} is the first magnitude estimate,
N is a number of samples, and
ν i is an i th input magnitude,
wherein the second magnitude estimate is generated using a formula:
v
^
=
Real
{
1
N
∑
i
v
i
2
-
2
σ
2
}
wherein
{circumflex over (ν)} is the second magnitude estimate,
N is a number of samples,
ν i is an i th input magnitude, and
σ is a standard deviation of noise in Gaussian distribution;
wherein the third magnitude estimate is generated using either one of a formula:
v
^
=
(
1
N
∑
i
v
i
)
(
1
-
e
1
N
∑
i
-
v
i
2
/
2
σ
2
A
)
or a quadrature specific formula:
v
^
=
(
Q
2
+
I
2
)
(
1
-
e
∑
i
-
(
Q
i
2
+
I
i
2
)
/
2
σ
2
A
)
wherein
{circumflex over (ν)} is the third magnitude estimate,
N is a number of samples,
ν i is an i th input magnitude, and
σ is a standard deviation of noise in Gaussian distribution,
A is a constant for a filter aggressiveness,
Q is quadrature components, and
I is in phase components; and
wherein the probability is reduced to:
p
(
v
>
0
|
Q
,
I
)
=
1
-
∏
i
e
-
Q
i
2
/
2
σ
2
A
∏
i
e
-
I
i
2
/
2
σ
2
A
wherein,
σ is a standard deviation of noise in Gaussian distribution, and
A is a constant for a filter aggressiveness.
4. The imaging system of claim 1 , wherein the second magnitude estimate and the third magnitude estimate are combined using arbitrary weights when averaging the second magnitude estimate and the third magnitude estimate to determine the fourth magnitude estimate.
5. The imaging system of claim 1 , wherein the second magnitude estimate and the third magnitude estimate are combined using calculated weights when averaging the second magnitude estimate and the third magnitude estimate to determine the fourth magnitude estimate.
6. The imaging system of claim 5 , wherein the calculated weights are determined based on error rates for true magnitude values.
7. The Imaging system of claim 1 , wherein the plurality of digital signals is generated by a technique comprising a Quadrature sampling, a Nyquist sampling, or any combination thereof.
8. The imaging system of claim 1 , wherein an input magnitude for each of the plurality of digital signals is determined by an algorithm comprising Euclidean-norm, Hilbert transform, envelope detection, demodulation, or any combination thereof.