U.S. patent application number 14/849391 was filed with the patent office on 2017-03-09 for parameter-free denoising of complex mr images by iterative multi-wavelet thresholding.
The applicant listed for this patent is Siemens Healthcare GmbH. Invention is credited to Stephan Kannengiesser, Boris Mailhe, Mariappan S. Nadar.
Application Number | 20170069082 14/849391 |
Document ID | / |
Family ID | 57964952 |
Filed Date | 2017-03-09 |
United States Patent
Application |
20170069082 |
Kind Code |
A1 |
Mailhe; Boris ; et
al. |
March 9, 2017 |
Parameter-Free Denoising of Complex MR Images by Iterative
Multi-Wavelet Thresholding
Abstract
A method for denoising Magnetic Resonance Imaging (MRI) data
includes receiving a noisy image acquired using an MRI imaging
device and determining a noise model comprising a non-diagonal
covariance matrix based on the noisy image and calibration
characteristics of the MRI imaging device. The noisy image is
designated as the current best image. Then, an iterative denoising
process is performed to remove noise from the noisy image. Each
iteration of the iterative denoising process comprises (i) applying
a bank of heterogeneous denoisers to the current best image to
generate a plurality of filter outputs, (ii) creating an image
matrix comprising the noisy image, the current best image, and the
plurality of filter outputs, (iii) finding a linear combination of
elements of the image matrix which minimizes a Stein Unbiased Risk
Estimation (SURE) value for the linear combination and the noise
model, (iv) designating the linear combination as the current best
image, and (v) updating each respective denoiser in the bank of
heterogeneous denoisers based on the SURE value. Following the
iterative denoising process, the current best image is designated
as a final denoised image.
Inventors: |
Mailhe; Boris; (Plainsboro,
NJ) ; Nadar; Mariappan S.; (Princeton, NJ) ;
Kannengiesser; Stephan; (Wuppertal, DE) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Siemens Healthcare GmbH |
Erlangen |
|
DE |
|
|
Family ID: |
57964952 |
Appl. No.: |
14/849391 |
Filed: |
September 9, 2015 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T 5/002 20130101;
G06T 5/10 20130101; G06T 2207/10088 20130101; G06T 5/20 20130101;
G01R 33/5611 20130101; G01R 33/5608 20130101 |
International
Class: |
G06T 7/00 20060101
G06T007/00; G06T 11/00 20060101 G06T011/00; G06T 5/10 20060101
G06T005/10; G06T 5/00 20060101 G06T005/00; G06T 5/20 20060101
G06T005/20 |
Claims
1. A method for denoising Magnetic Resonance Imaging (MRI) data,
the method comprising: receiving a noisy image acquired using an
MRI imaging device; determining a noise model comprising a
non-diagonal covariance matrix based on the noisy image and
calibration characteristics of the MRI imaging device; designating
the noisy image as a current best image; performing an iterative
denoising process to remove noise from the noisy image, each
iteration of the iterative denoising process comprising: applying a
bank of heterogeneous denoisers to the current best image to
generate a plurality of filter outputs, creating an image matrix
comprising the noisy image, the current best image, and the
plurality of filter outputs, finding a linear combination of
elements of the image matrix which minimizes a Stein Unbiased Risk
Estimation (SURE) value for the linear combination and the noise
model, designating the linear combination as the current best
image, updating each respective denoiser in the bank of
heterogeneous denoisers based on the SURE value; and following the
iterative denoising process, designating the current best image as
a final denoised image.
2. The method of claim 1, wherein each respective denoiser in the
bank of heterogeneous denoisers corresponds to a wavelet
filter.
3. The method of claim 2, wherein each respective denoiser in the
bank of heterogeneous denoisers corresponds to a distinct
Daubechies wavelet.
4. The method of claim 2, wherein the iterative denoising process
further comprising: performing a sequential cycle spinning
operation to shift transforms performed by each respective wavelet
filter by a predetermined offset value.
5. The method of claim 4, wherein the predetermined offset value
varies during each iteration of the iterative denoising
process.
6. The method of claim 1, wherein each respective denoiser in the
bank of heterogeneous denoisers uses a thresholding operator
corresponding to an empirical Wiener filter.
7. The method of claim 6, wherein the thresholding operator applies
non-negative garrote thresholding.
8. The method of claim 6, wherein updating each respective denoiser
in the bank of heterogeneous denoisers based on the SURE value
comprises: updating the thresholding operator used by each
respective denoiser based on the SURE value.
9. The method of claim 1, further comprising: acquiring k-space
data using the MRI imaging device; determining a spatial modulation
measurement operator corresponding to the MRI imaging device based
on the k-space data; applying one or more filters to the k-space
data to yield filtered k-space data and a spectral measurement
operator; reconstructing the noisy image based on the filtered
k-space data; generating the noise model based on the spatial
modulation measurement operator and the spectral measurement
operator.
10. The method of claim 1, wherein one or more denoisers included
in the bank of heterogeneous denoisers is configured based on the
noise model.
11. The method of claim 1, wherein the iterative denoising process
further comprises: selecting a subset of the plurality of filter
outputs, wherein each respective output in the subset corresponds
to a recombination weight in the linear combination below a
predetermined threshold value; identifying one or more low-weighted
denoisers in the bank of heterogeneous denoisers corresponding to
the plurality of filter outputs; and removing the one or more
low-weighted denoisers from the bank of heterogeneous
denoisers.
12. An article of manufacture for denoising Magnetic Resonance
Imaging (MRI) data, the article of manufacture comprising a
non-transitory, tangible computer-readable medium holding
computer-executable instructions for performing a method
comprising: receiving a noisy image acquired using an MRI imaging
device; determining a noise model comprising a non-diagonal
covariance matrix based on the noisy image and calibration
characteristics of the MRI imaging device; designating the noisy
image as a current best image; performing an iterative denoising
process to remove noise from the noisy image, each iteration of the
iterative denoising process comprising: applying a bank of
heterogeneous denoisers to the current best image to generate a
plurality of filter outputs, creating an image matrix comprising
the noisy image, the current best image, and the plurality of
filter outputs, finding a linear combination of elements of the
image matrix which minimizes a Stein Unbiased Risk Estimation
(SURE) value for the linear combination and the noise model,
designating the linear combination as the current best image,
updating each respective denoiser in the bank of heterogeneous
denoisers based on the SURE value; and following the iterative
denoising process, designating the current best image as a final
denoised image.
13. The article of manufacture of claim 12, wherein each respective
denoiser in the bank of heterogeneous denoisers corresponds to a
wavelet filter.
14. The article of manufacture of claim 13, wherein the iterative
denoising process further comprising: performing a sequential cycle
spinning operation to shift transforms performed by each respective
wavelet filter by a predetermined offset value that varies during
each iteration of the iterative denoising process.
15. The article of manufacture of claim 12, wherein each respective
denoiser in the bank of heterogeneous denoisers uses a thresholding
operator corresponding to an empirical Wiener filter.
16. The article of manufacture of claim 15, wherein the
thresholding operator applies non-negative garrote
thresholding.
17. The article of manufacture of claim 15, wherein updating each
respective denoiser in the bank of heterogeneous denoisers based on
the SURE value comprises: updating the thresholding operator used
by each respective denoiser based on the SURE value.
18. The article of manufacture of claim 12, wherein the method
further comprises: receiving k-space data acquired using the MRI
imaging device; determining a spatial modulation measurement
operator corresponding to the MRI imaging device based on the
k-space data; applying one or more filters to the k-space data to
yield filtered k-space data and a spectral measurement operator;
reconstructing the noisy image based on the filtered k-space data;
generating the noise model based on the spatial modulation
measurement operator and the spectral measurement operator.
19. The article of manufacture of claim 12, wherein one or more
denoisers included in the bank of heterogeneous denoisers is
configured based on the noise model.
20. A system for denoising Magnetic Resonance Imaging (MRI) data,
the system comprising: an MRI imaging device configured to acquire
k-space data; a computer comprising an image data processor
configured to: reconstruct a noisy image based on the k-space data,
determine a noise model comprising a non-diagonal covariance matrix
based on the noisy image and calibration characteristics of the MRI
imaging device; designate the noisy image as a current best image;
perform an iterative denoising process to remove noise from the
noisy image, each iteration of the iterative denoising process
comprising: (i) applying a bank of heterogeneous denoisers to the
current best image to generate a plurality of filter outputs, (ii)
creating an image matrix comprising the noisy image, the current
best image, and the plurality of filter outputs, (iii) finding a
linear combination of elements of the image matrix which minimizes
a Stein Unbiased Risk Estimation (SURE) value for the linear
combination and the noise model, (iv) designating the linear
combination as the current best image, (v) updating each respective
denoiser in the bank of heterogeneous denoisers based on the SURE
value; and following the iterative denoising process, designate the
current best image as a final denoised image.
Description
TECHNOLOGY FIELD
[0001] The present invention relates generally to methods, systems,
and apparatuses for performing Magnetic Resonance Imaging (MRI)
reconstruction using a parameter-free denoising technique that is
applied to iterative multi-wavelet thresholding to complex MRI
data. The disclosed techniques may be applied to enhance and
simplify existing denoising MRI operations.
BACKGROUND
[0002] MR images are acquired in k-space in the Fourier domain. The
acquired image data includes white noise. That is, noise in the
acquired image data is a signal whose spectrum has equal power
within any equal interval of frequencies. In order to transform the
raw image data into an image, a reconstruction process is
performed. Using a standard linear reconstruction model, one or
more filtering operations must be performed during the
reconstruction. During each filtering operation, multiplications
may be performed to color the noise by making certain frequencies
of the noise signal more prominent than others. This creates a
modulation of the noise in the Fourier domain.
[0003] Following filtering, a coil combination process is performed
to combine data acquired from individual coils of the MRI device.
Coil combination may also encompass so-called "parallel imaging",
in which undersampling artifacts through accelerated acquisition
are compensated for via dedicated coil combination coefficients.
This combination may also amplify the noise in some region of the
space, in the pixel domain. This noise "modulation" is often
referred to as "G-factor" noise and may be influenced by the coil
geometry, the field of view, the location of the imaging plane, and
the sampling pattern used to acquire the k-space data. This creates
modulation of the noise in space.
[0004] An additional modulation in the noise results from the
signal level not being uniform in space. During imaging, a coil
array is placed around the subject's body. Signal sensitivity is
typically very good close to the coils (i.e., at the periphery of
the body); however, sensitivity is poor in the center of image
space. So, if an image is reconstructed based on this data, it will
have a dark center region. To compensate for the difference in
sensitivity, a corrective field may be applied in the pixel domain
(i.e., multiply the pixel intensity by some factor in the center);
however, this modulates the noise as well.
[0005] Accordingly, the noise associated with an MRI acquisition is
Gaussian, but modulated in both in the Fourier domain and pixel
domain. Conventional denoising techniques are unable to address
denoising in this context because these techniques cannot
adequately handle noise in the Fourier domain and, at the same
time, adapt to varying noise level so denoising of the center of
the image does not result in blurring.
SUMMARY
[0006] Embodiments of the present invention address and overcome
one or more of the above shortcomings and drawbacks, by providing
methods, systems, and apparatuses related to a self-tuning
iterative denoising algorithm for Magnetic Resonance Imaging (MRI)
applications that adjusts its parameters based on the observed
image and noise parameters gathered at acquisition.
[0007] According to some embodiments of the present invention, a
method for denoising MRI data includes receiving a noisy image
acquired using an MRI imaging device and determining a noise model
comprising a non-diagonal covariance matrix based on the noisy
image and calibration characteristics of the MRI imaging device.
The noisy image is designated as the current best image. Then, an
iterative denoising process is performed to remove noise from the
noisy image. Each iteration of the iterative denoising process
comprises (i) applying a bank of heterogeneous denoisers to the
current best image to generate a plurality of filter outputs, (ii)
creating an image matrix comprising the noisy image, the current
best image, and the plurality of filter outputs, (iii) finding a
linear combination of elements of the image matrix which minimizes
a Stein Unbiased Risk Estimation (SURE) value for the linear
combination and the noise model, (iv) designating the linear
combination as the current best image, and (v) updating each
respective denoiser in the bank of heterogeneous denoisers based on
the SURE value. Following the iterative denoising process, the
current best image is designated as a final denoised image.
[0008] The features of the aforementioned method can be refined,
supplemented, or otherwise modified in different embodiments of the
present invention. In some embodiments, each respective denoiser in
the bank of heterogeneous denoisers corresponds to a wavelet
filter. For example, in some embodiments, each respective denoiser
corresponds to a distinct Daubechies wavelet. In some embodiments,
a sequential cycle spinning operation is performed during the
iterative denoising process to shift transforms performed by each
respective wavelet filter by a predetermined offset value that may
vary during each iteration. In some embodiments, one or more
denoisers included in the bank of heterogeneous denoisers is
configured based on the noise model.
[0009] In some embodiments of the aforementioned method, each
respective denoiser in the bank of heterogeneous denoisers uses a
thresholding operator corresponding to an empirical Wiener filter.
For example, in one embodiment, the thresholding operator applies
non-negative garrote thresholding. During the aforementioned
method, each respective denoiser in the bank of heterogeneous
denoisers may be updated by updating the thresholding operator used
by each respective denoiser based on the SURE value.
[0010] In some embodiments of the aforementioned method, k-space
data is acquired using the MRI imaging device. Next, a spatial
modulation measurement operator corresponding to the MRI imaging
device is determined based on the k-space data. One or more filters
are applied to the k-space data to yield filtered k-space data and
a spectral measurement operator. The noisy image is reconstructed
based on the filtered k-space data. Then, the noise model is
generated based on the spatial modulation measurement operator and
the spectral measurement operator.
[0011] In other embodiments of the aforementioned method, the
iterative denoising process further comprises selecting a subset of
the filter outputs, where each respective output in the subset
corresponds to a recombination weight in the linear combination
below a predetermined threshold value. Next, low-weighted denoisers
in the bank of heterogeneous denoisers corresponding to the filter
outputs are identified. Then, the low-weighted denoisers are
removed from the bank of heterogeneous denoisers.
[0012] According to other embodiments of the present invention, an
article of manufacture for denoising MRI data comprises a
non-transitory, tangible computer-readable medium holding
computer-executable instructions for performing the aforementioned
method, with or without the additional features discussed
above.
[0013] According to other embodiments, a system for denoising MRI
data comprises an MRI imaging device configured to acquire k-space
data and a computer comprising an image data processor. The image
data processor is configured to reconstruct a noisy image based on
the k-space data, determine a noise model comprising a non-diagonal
covariance matrix based on the noisy image and calibration
characteristics of the MRI imaging device, designate the noisy
image as a current best image, and perform an iterative denoising
process to remove noise from the noisy image. Each iteration of the
iterative denoising process comprises (i) applying a bank of
heterogeneous denoisers to the current best image to generate a
plurality of filter outputs, (ii) creating an image matrix
comprising the noisy image, the current best image, and the
plurality of filter outputs, (iii) finding a linear combination of
elements of the image matrix which minimizes a SURE value for the
linear combination and the noise model, (iv) designating the linear
combination as the current best image, and (v) updating each
respective denoiser in the bank of heterogeneous denoisers based on
the SURE value. Following the iterative denoising process, the
image data processor designates the current best image as a final
denoised image.
[0014] Additional features and advantages of the invention will be
made apparent from the following detailed description of
illustrative embodiments that proceeds with reference to the
accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The foregoing and other aspects of the present invention are
best understood from the following detailed description when read
in connection with the accompanying drawings. For the purpose of
illustrating the invention, there is shown in the drawings
embodiments that are presently preferred, it being understood,
however, that the invention is not limited to the specific
instrumentalities disclosed. Included in the drawings are the
following Figures:
[0016] FIG. 1 shows a system for ordering acquisition of frequency
domain components representing magnetic resonance image data for
storage in a k-space storage array, as used by some embodiments of
the present invention;
[0017] FIG. 2 illustrates a process for generating a noisy image
and a noise model during MRI acquisition, according to some
embodiments;
[0018] FIG. 3 provides an overview of an iterative denoising
process that may be applied to denoise an image, according to some
embodiments;
[0019] FIG. 4A shows two images acquired using an MP2RAGE sequence
for two different contrast levels where no denoising is
applied;
[0020] FIG. 4B shows the two denoised image corresponding to the
images shown in FIG. 4A, where the denoising is performed using the
techniques described herein; and
[0021] FIG. 5 illustrates an exemplary computing environment within
which embodiments of the invention may be implemented.
DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS
[0022] The following disclosure describes the present invention
according to several embodiments directed at methods, systems, and
apparatuses related to performing Magnetic Resonance Imaging (MRI)
reconstruction using a parameter-free denoising technique that
apply iterative multi-wavelet thresholding to complex MRI data.
Briefly, the techniques described herein use a self-tuning
iterative denoising algorithm that adjusts its parameters based on
the observed image and noise parameters gathered at acquisition.
During each iteration, the current iterate is filtered by a bank of
wavelet denoisers, using an empirical Wiener filter to adapt to the
noise level. Then, the different filter outputs are recombined
among each other, as well as with the previous iterate and the
original image to adjust the global filtering strength. The
recombination weights may be computed on the fly to minimize
Stein's Unbiased Risk Estimation (SURE), which allows the algorithm
to adapt to the image and noise contents. The SURE value itself
also provides an estimate of the residual noise variance in the
current iterate, which can be used in the next iteration to adjust
the thresholds. To reduce computational costs, decimated wavelets
may be used in the individual denoisers, but with different shifts
at each iteration so that the overall denoising benefits from the
whole undecimated transform.
[0023] The iterative denoising process includes several
developments which distinguish it from conventional denoising
techniques in order to efficiently apply SURE to the denoising of
MRI images. For example, because the images being denoised are
complex, the SURE linear expansion is expanded to complex
coefficients by considering the real and imaginary parts
separately. Additionally, the noise in MR images is modulated not
only in voxel domain, but also in k-space due to filtering steps
applied before reconstruction. This results in a noise model where
the axes of the noise covariance may be unknown. To address this,
SURE is extended to manipulate covariance operators in an abstract
way instead of relying on a diagonal form.
[0024] FIG. 1 shows a system 100 for ordering acquisition of
frequency domain components representing MRI data for storage in a
k-space storage array, as used by some embodiments of the present
invention. This system 100 is sometimes referred to herein
collectively as the "MRI Scanner." In system 100, magnetic coils 12
create a static base magnetic field in the body of patient 11 to be
imaged and positioned on a table. Within the magnet system are
gradient coils 14 for producing position dependent magnetic field
gradients superimposed on the static magnetic field. Gradient coils
14, in response to gradient signals supplied thereto by a gradient
and shim coil control module 16, produce position dependent and
shimmed magnetic field gradients in three orthogonal directions and
generates magnetic field pulse sequences. The shimmed gradients
compensate for inhomogeneity and variability in an MRI device
magnetic field resulting from patient anatomical variation and
other sources. The magnetic field gradients include a
slice-selection gradient magnetic field, a phase-encoding gradient
magnetic field and a readout gradient magnetic field that are
applied to patient 11.
[0025] Further radio frequency (RF) module 20 provides RF pulse
signals to RF coil 18, which in response produces magnetic field
pulses which rotate the spins of the protons in the imaged body of
the patient 11 by ninety degrees or by one hundred and eighty
degrees for so-called "spin echo" imaging, or by angles less than
or equal to 90 degrees for so-called "gradient echo" imaging.
Gradient and shim coil control module 16 in conjunction with RF
module 20, as directed by central control unit 26, control
slice-selection, phase-encoding, readout gradient magnetic fields,
radio frequency transmission, and magnetic resonance signal
detection, to acquire magnetic resonance signals representing
planar slices of patient 11.
[0026] In response to applied RF pulse signals, the RF coil 18
receives magnetic resonance signals, i.e., signals from the excited
protons within the body as they return to an equilibrium position
established by the static and gradient magnetic fields. The
magnetic resonance signals are detected and processed by a detector
within RF module 20 and k-space component processor unit 34 to
provide a magnetic resonance dataset to an image data processor for
processing into an image. In some embodiments, the image data
processor is located in central control unit 26. However, in other
embodiments such as the one depicted in FIG. 1, the image data
processor is located in a separate unit 27. ECG synchronization
signal generator 30 provides ECG signals used for pulse sequence
and imaging synchronization. A two or three dimensional k-space
storage array of individual data elements in k-space component
processor unit 34 stores corresponding individual frequency
components comprising a magnetic resonance dataset. The k-space
array of individual data elements has a designated center and
individual data elements individually have a radius to the
designated center.
[0027] A magnetic field generator (comprising coils 12, 14, and 18)
generates a magnetic field for use in acquiring multiple individual
frequency components corresponding to individual data elements in
the storage array. The individual frequency components are
successively acquired in an order in which radius of respective
corresponding individual data elements increases and decreases
along a substantially spiral path as the multiple individual
frequency components is sequentially acquired during acquisition of
a magnetic resonance dataset representing an magnetic resonance
image. A storage processor in the k-space component processor unit
34 stores individual frequency components acquired using the
magnetic field in corresponding individual data elements in the
array. The radius of respective corresponding individual data
elements alternately increases and decreases as multiple sequential
individual frequency components are acquired. The magnetic field
acquires individual frequency components in an order corresponding
to a sequence of substantially adjacent individual data elements in
the array and magnetic field gradient change between successively
acquired frequency components is substantially minimized.
[0028] Central control unit 26 uses information stored in an
internal database to process the detected magnetic resonance
signals in a coordinated manner to generate high quality images of
a selected slice(s) of the body (e.g., using the image data
processor) and adjusts other parameters of system 100. The stored
information comprises predetermined pulse sequence and magnetic
field gradient and strength data as well as data indicating timing,
orientation and spatial volume of gradient magnetic fields to be
applied in imaging. Generated images are presented on display 40 of
the operator interface. Computer 28 of the operator interface
includes a graphical user interface (GUI) enabling user interaction
with central control unit 26 and enables user modification of
magnetic resonance imaging signals in substantially real time.
Continuing with reference to FIG. 1, display processor 37 processes
the magnetic resonance signals to reconstruct one or more images
for presentation on display 40, for example. Various techniques may
be used for reconstruction. For example, as described in greater
detail below, an optimization algorithm is applied to iteratively
solve a cost function which results in the reconstructed image.
[0029] FIG. 2 illustrates a process 200 for generating a Noisy
Image 250 and a Noise Model 260 during MRI acquisition, according
to some embodiments. K-Space Data 205 is acquired, as described
above with reference to FIG. 1. At the MRI Scanner an
Autocalibration Process 210 is performed. The Autocalibration
Process 210 produces Generalized Autocalibrating Partially Parallel
Acquisition (GRAPPA) Kernels 220. As is generally understood in the
art, GRAPPA is a parallel imaging technique to speed up MRI pulse
sequences. GRAPPA introduces G-Factor noise modulation in the voxel
space. This noise modulation is captured by the Autocalibration
Process 210 and represented in FIG. 2 by Spatial Modulation
Measurement 215. Additionally, a residual bias field may also be
captured in the Spatial Modulation Measurement 215.
[0030] Continuing with reference to FIG. 2, a Filtering Process 225
is also applied to K-Space Data 205 to produce Filtered K-Space
Data 235. Using a standard linear reconstruction model, the
Filtering Process 225 performs multiplications which color the
noise in the K-Space Data 205 by making certain frequencies of the
noise signal more prominent than others. This creates a modulation
of the noise in the Fourier domain which is represented in FIG. 2
by Spectral Modulation Measurement 230.
[0031] Using the GRAPPA Kernels 220 and the Filtered K-Space Data
235, a Reconstruction Process 240 may be performed to result in a
Noisy Image 250. Notably, at this point, little or no denoising may
have been performed on the data. However, the Spatial Modulation
Measurement 215 and the Spectral Modulation Measurement 230 may be
used to create a Noise Model 260 which, as described below in
reference to FIG. 3, may be used to denoise the image.
X.about.(X.sub.0,.SIGMA..sup.2) (1)
.SIGMA..sup.2=.DELTA..sub.pF.sup.-1.DELTA..sub.k.sup.2F.DELTA..sub.p
(2)
In these equations F is the Fourier transform, .DELTA..sub.k and
.DELTA..sub.p represent the Spatial Modulation Measurement 215 and
the Spectral Modulation Measurement 230, respectively. X is the
signal (i.e., images) and X.sub.0 is the ground truth (i.e., the
denoised image). .SIGMA..sup.2 is the covariance of the noise.
[0032] FIG. 3 provides an overview of an iterative denoising
process 300 that may be applied to denoise an image, according to
some embodiments. This process 300 may be implemented, for example,
by a computer using the image data processor described above with
reference to FIG. 1. The initial inputs to the process are the
Noisy Image 355 and the Initial Noise Model 360. These two items
may be generated using the process 200 described above with
reference to FIG. 2. The Noisy Image 355 and the Initial Noise
Model 360 serve as the Current Image 305 and the Noise Model 310
for the first iteration of the iterative denoising process 200.
[0033] During each iteration of the iterative denoising process
300, Denoisers 315, 320, 325, 330 are applied to the Current Image
305 and the Noise Model 310. It should be noted that, although four
denoisers are shown in FIG. 3, a greater or smaller number of
denoisers may be used in other embodiments. The Denoisers 315, 320,
325, 330 are "atomic" in the sense that they operate
independently.
[0034] Any type of denoiser may be used in the process 300 shown in
FIG. 3. For example, in some embodiments, the Denoisers 315, 320,
325, 330 are simple orthogonal wavelet thresholding filters where
wavelet coefficient values not exceeding a selected threshold level
are replaced by zero followed by an inverse transform of modified
coefficients to recover denoised signal. This may be represented as
follows:
x.rarw.W.sup.TS.sub..sigma.Wx (3)
In Equation 3, W is a wavelet transform with three levels. For
example, in one embodiment, the denoising filters are Daubechies 1
to 8. Each Denoiser 315, 320, 325, 330 denoises the image in a
slightly different way. Additionally, the wavelet applied by each
of the Denoisers 315, 320, 325, 330 introduces some artifacts. For
example, for a Haar wavelet filter, the filtered image may include
some "staircasing" such that it appears piece-wise constant. With
higher order wavelet filters there may be some "ringing" where
there are ripples around certain edges in the image. By using
different filters, these different artifacts may be cancelled out,
thus resulting in an image with less overall noise. S in Equation 3
is a non-negative garrote thresholding operator that corresponds to
an empirical Wiener filter. The parameter may be determined
directly from the noise level as follows:
S .sigma. ( x ) = max ( x 2 - .sigma. 2 , 0 ) x 2 x ( 4 )
##EQU00001##
Here, .sigma..sup.2 is the variance of the noise level. In some
embodiments, sequential cycle spinning is used to shift the
transforms performed by the filters by a different offset at each
iteration. This provides the quality of redundant transforms for
the cost of orthogonal ones and avoids blocking artifacts.
[0035] The results generated by Denoisers 315, 320, 325, 330 are
combined via a Stein Unbiased Risk Estimation (SURE) Component.
SURE is an estimation technique generally known in the art and,
thus, only briefly described here. Let x(x.sub.0,.sigma..sup.2) be
a noisy observation of x.sub.0 and let h(x) be a denoiser (i.e., an
estimator of x.sub.0, given x). The quality of h can be measured by
quadratic risk as follows:
R(h)=E(.parallel.h(x)-x.sub.0.parallel..sup.2). However, in
general, x.sub.0 is unknown and the risk can only be computed on
synthetic data. SURE avoids computation of x.sub.0 by providing an
estimator of risk that only depends on x. Continuing with the
notation set out above, the general form of SURE may be expressed
by:
SURE ( h , x ) = h ( x ) - x 2 - .sigma. 2 + 2 .sigma. 2
.differential. h .differential. x ( x ) ( 5 ) ##EQU00002##
This may be generalized to dimension N with white Gaussian noise
(or diagonal covariance) as follows:
SURE(H,X)=.parallel.H(X)-X.parallel..sup.2-N.sigma..sup.2+2.sigma..sup.2
div H(x) (6)
[0036] In MRI applications, the noise is not white and must be
represented by a non-diagonal covariance. For such applications,
the generalization described in Equation 6 is not applicable.
Rather, Equation 5 is generalized as follows:
SURE(H,X)=.parallel.H(X)-X.parallel..sup.2-Tr(.SIGMA..sup.2)+2 div
.SIGMA..sup.2H(x) (7)
Equation 7 may further be generalized to complex numbers (such as
available in MRI applications) by considering each complex variable
as a pair of real variables with independent and identically
distributed (i.i.d.) noise.
[0037] SURE-LET is the accepted name of SURE applied to a linear
model combination. For atomic denoisers (such as Denoisers 315,
320, 325, 330), one gets n images plus the noisy one and the
previous iterate: Y=[X.sub.1 . . . X.sub.N]. Based on this data,
one may determine the recombination weights that minimize SURE via
a simple low-dimension least-square problem as follows:
.lamda. ^ = argmin .lamda. Y .lamda. - X 2 - Tr ( .SIGMA. 2 ) + 2
div .SIGMA. 2 Y .lamda. ( x ) = argmin .lamda. Y .lamda. - X 2 - Tr
( .SIGMA. 2 ) + 2 .lamda. , [ div .SIGMA. 2 X n ( x ) ] 1 .ltoreq.
n .ltoreq. N ( 9 ) = ( Y T Y ) - 1 ( Y T X - [ div .SIGMA. 2 X n (
X ) ] 1 .ltoreq. n .ltoreq. N ) ( 10 ) ( 8 ) ##EQU00003##
To simplify matters further, in some embodiments, these equations
may be constrained, for example, by limiting the computations to
real weight values and/or only positive weight values.
[0038] Closed-form derivation of Equation 10 would require
differentiation of the denoising process with respect to the input
image, which is intractable for iterative methods as the chain rule
introduces large-scale spatial dependencies. Instead, the SURE
Component 340 may perform the SURE Monte Carlo method where
derivatives are approximated by the difference between the
denoising of the actual image and an image with slightly more noise
added:
.DELTA. i . i . d . ~ ( 0 , 1 ) ( 11 ) .epsilon. 1 ( 12 ) div
.SIGMA. 2 H ( X ) .apprxeq. 1 .epsilon. .SIGMA. 2 H ( X +
.epsilon..DELTA. ) - .SIGMA. 2 H ( X ) , .DELTA. ( 13 )
##EQU00004##
To track the risk over iterations, .DELTA. is drawn once at the
beginning of the algorithm and the images X and X+.epsilon..DELTA.
are denoised simultaneously. The overhead cost is compensated by
the adaptive step size that ensures optimal convergence speed of
the algorithm.
[0039] The SURE Component 340 generates a denoised image 350 which
may be used as the Current Image 305 for the next iterate.
Additionally, the SURE Component 340 provides a measure of estimate
risk (i.e., the remaining noise level in the next iterate). This
information is used to refine the Noise Model 310 for use in the
next iteration.
[0040] The iterative denoising process 300 repeats for a
predetermined number of iterations (e.g., supplied by a user) or
until some threshold based criteria is reached (e.g., estimate risk
below a certain value). In some embodiments, the weight values
provided by SURE may be used to reduce the number of filters used
for later iterations. For example, all filters with weight values
below a threshold value may be eliminated from further process.
This helps to accelerate the overall processing time of the
iterative denoising process 300. Once the iterative denoising
process 300 is complete, the denoised image 350 is the final
denoised image.
[0041] The iterative denoising process 300 described in FIG. 3
operates on complex data. Thus, the Denoisers 315, 320, 325, 330
and the SURE Component 340 are able to leverage phase information
and avoid Rician effects. It should also be noted that the
iterative denoising process 300 is self-tuning. Empirical Wiener
filtering matches thresholds to local noise levels and SURE
recombination adapts weights of different filters and algorithm
step size. The result is uniform over space and, because the noise
model is gathered at the scanner (see FIG. 2), no user input is
required. Additionally, because multiple filters are used, the
iterative denoising process 300 provides the benefits of
multi-model fusion (e.g., minimizing regularization artifacts).
[0042] FIGS. 4A and 4B show a set of images which demonstrate the
benefits of denoising using the techniques described above with
respect to FIGS. 2 and 3. More specifically, FIG. 4A shows two
images 405, 410 acquired using an MP2RAGE sequence for two
different contrast levels where no denoising is applied. FIG. 4B
shows two images 415, 420 demonstrating the effect of applying the
iterative denoising process described herein. Note that the signal
and noise of images can vary greatly from one contrast level to the
next. Thus, the denoising effects shown in FIG. 4B demonstrates
that the iterative denoising process is robust such that it can
adapt to the changes and still provide high-quality denoising.
[0043] FIG. 5 illustrates an exemplary computing environment 500
within which embodiments of the invention may be implemented. For
example, this computing environment 500 may be used to implement
the processes 200, 300, described in FIGS. 2, and 3, respectively.
In some embodiments, the computing environment 500 may be used to
implement one or more of the components illustrated in the system
100 of FIG. 1. The computing environment 500 may include computer
system 510, which is one example of a computing system upon which
embodiments of the invention may be implemented. Computers and
computing environments, such as computer system 510 and computing
environment 500, are known to those of skill in the art and thus
are described briefly here.
[0044] As shown in FIG. 5, the computer system 510 may include a
communication mechanism such as a bus 521 or other communication
mechanism for communicating information within the computer system
510. The computer system 510 further includes one or more
processors 520 coupled with the bus 521 for processing the
information. The processors 520 may include one or more central
processing units (CPUs), graphical processing units (GPUs), or any
other processor known in the art.
[0045] The computer system 510 also includes a system memory 530
coupled to the bus 521 for storing information and instructions to
be executed by processors 520. The system memory 530 may include
computer readable storage media in the form of volatile and/or
nonvolatile memory, such as read only memory (ROM) 531 and/or
random access memory (RAM) 532. The system memory RAM 532 may
include other dynamic storage device(s) (e.g., dynamic RAM, static
RAM, and synchronous DRAM). The system memory ROM 531 may include
other static storage device(s) (e.g., programmable ROM, erasable
PROM, and electrically erasable PROM). In addition, the system
memory 530 may be used for storing temporary variables or other
intermediate information during the execution of instructions by
the processors 520. A basic input/output system (BIOS) 533
containing the basic routines that help to transfer information
between elements within computer system 510, such as during
start-up, may be stored in ROM 531. RAM 532 may contain data and/or
program modules that are immediately accessible to and/or presently
being operated on by the processors 520. System memory 530 may
additionally include, for example, operating system 534,
application programs 535, other program modules 536 and program
data 537.
[0046] The computer system 510 also includes a disk controller 540
coupled to the bus 521 to control one or more storage devices for
storing information and instructions, such as a hard disk 541 and a
removable media drive 542 (e.g., floppy disk drive, compact disc
drive, tape drive, and/or solid state drive). The storage devices
may be added to the computer system 510 using an appropriate device
interface (e.g., a small computer system interface (SCSI),
integrated device electronics (IDE), Universal Serial Bus (USB), or
FireWire).
[0047] The computer system 510 may also include a display
controller 565 coupled to the bus 521 to control a display 566,
such as a cathode ray tube (CRT) or liquid crystal display (LCD),
for displaying information to a computer user. The computer system
includes an input interface 560 and one or more input devices, such
as a keyboard 562 and a pointing device 561, for interacting with a
computer user and providing information to the processor 520. The
pointing device 561, for example, may be a mouse, a trackball, or a
pointing stick for communicating direction information and command
selections to the processor 520 and for controlling cursor movement
on the display 566. The display 566 may provide a touch screen
interface which allows input to supplement or replace the
communication of direction information and command selections by
the pointing device 561.
[0048] The computer system 510 may perform a portion or all of the
processing steps of embodiments of the invention in response to the
processors 520 executing one or more sequences of one or more
instructions contained in a memory, such as the system memory 530.
Such instructions may be read into the system memory 530 from
another computer readable medium, such as a hard disk 541 or a
removable media drive 542. The hard disk 541 may contain one or
more datastores and data files used by embodiments of the present
invention. Datastore contents and data files may be encrypted to
improve security. The processors 520 may also be employed in a
multi-processing arrangement to execute the one or more sequences
of instructions contained in system memory 530. In alternative
embodiments, hard-wired circuitry may be used in place of or in
combination with software instructions. Thus, embodiments are not
limited to any specific combination of hardware circuitry and
software.
[0049] As stated above, the computer system 510 may include at
least one computer readable medium or memory for holding
instructions programmed according to embodiments of the invention
and for containing data structures, tables, records, or other data
described herein. The term "computer readable medium" as used
herein refers to any medium that participates in providing
instructions to the processor 520 for execution. A computer
readable medium may take many forms including, but not limited to,
non-volatile media, volatile media, and transmission media.
Non-limiting examples of non-volatile media include optical disks,
solid state drives, magnetic disks, and magneto-optical disks, such
as hard disk 541 or removable media drive 542. Non-limiting
examples of volatile media include dynamic memory, such as system
memory 530. Non-limiting examples of transmission media include
coaxial cables, copper wire, and fiber optics, including the wires
that make up the bus 521. Transmission media may also take the form
of acoustic or light waves, such as those generated during radio
wave and infrared data communications.
[0050] The computing environment 500 may further include the
computer system 510 operating in a networked environment using
logical connections to one or more remote computers, such as remote
computer 580. Remote computer 580 may be a personal computer
(laptop or desktop), a mobile device, a server, a router, a network
PC, a peer device or other common network node, and typically
includes many or all of the elements described above relative to
computer system 510. When used in a networking environment,
computer system 510 may include modem 572 for establishing
communications over a network 571, such as the Internet. Modem 572
may be connected to bus 521 via user network interface 570, or via
another appropriate mechanism.
[0051] Network 571 may be any network or system generally known in
the art, including the Internet, an intranet, a local area network
(LAN), a wide area network (WAN), a metropolitan area network
(MAN), a direct connection or series of connections, a cellular
telephone network, or any other network or medium capable of
facilitating communication between computer system 510 and other
computers (e.g., remote computer 580). The network 571 may be
wired, wireless or a combination thereof. Wired connections may be
implemented using Ethernet, Universal Serial Bus (USB), RJ-11 or
any other wired connection generally known in the art. Wireless
connections may be implemented using Wi-Fi, WiMAX, and Bluetooth,
infrared, cellular networks, satellite or any other wireless
connection methodology generally known in the art. Additionally,
several networks may work alone or in communication with each other
to facilitate communication in the network 571.
[0052] The embodiments of the present disclosure may be implemented
with any combination of hardware and software. In addition, the
embodiments of the present disclosure may be included in an article
of manufacture (e.g., one or more computer program products)
having, for example, computer-readable, non-transitory media. The
media has embodied therein, for instance, computer readable program
code for providing and facilitating the mechanisms of the
embodiments of the present disclosure. The article of manufacture
can be included as part of a computer system or sold
separately.
[0053] While various aspects and embodiments have been disclosed
herein, other aspects and embodiments will be apparent to those
skilled in the art. The various aspects and embodiments disclosed
herein are for purposes of illustration and are not intended to be
limiting, with the true scope and spirit being indicated by the
following claims.
[0054] An executable application, as used herein, comprises code or
machine readable instructions for conditioning the processor to
implement predetermined functions, such as those of an operating
system, a context data acquisition system or other information
processing system, for example, in response to user command or
input. An executable procedure is a segment of code or machine
readable instruction, sub-routine, or other distinct section of
code or portion of an executable application for performing one or
more particular processes. These processes may include receiving
input data and/or parameters, performing operations on received
input data and/or performing functions in response to received
input parameters, and providing resulting output data and/or
parameters.
[0055] A graphical user interface (GUI), as used herein, comprises
one or more display images, generated by a display processor and
enabling user interaction with a processor or other device and
associated data acquisition and processing functions. The GUI also
includes an executable procedure or executable application. The
executable procedure or executable application conditions the
display processor to generate signals representing the GUI display
images. These signals are supplied to a display device which
displays the image for viewing by the user. The processor, under
control of an executable procedure or executable application,
manipulates the GUI display images in response to signals received
from the input devices. In this way, the user may interact with the
display image using the input devices, enabling user interaction
with the processor or other device.
[0056] The functions and process steps herein may be performed
automatically or wholly or partially in response to user command.
An activity (including a step) performed automatically is performed
in response to one or more executable instructions or device
operation without user direct initiation of the activity.
[0057] The system and processes of the figures are not exclusive.
Other systems, processes and menus may be derived in accordance
with the principles of the invention to accomplish the same
objectives. Although this invention has been described with
reference to particular embodiments, it is to be understood that
the embodiments and variations shown and described herein are for
illustration purposes only. Modifications to the current design may
be implemented by those skilled in the art, without departing from
the scope of the invention. As described herein, the various
systems, subsystems, agents, managers and processes can be
implemented using hardware components, software components, and/or
combinations thereof. No claim element herein is to be construed
under the provisions of 35 U.S.C. 112, sixth paragraph, unless the
element is expressly recited using the phrase "means for."
* * * * *