U.S. patent application number 11/320743 was filed with the patent office on 2006-06-08 for magnetic resonance imaging with resolution and contrast enhancement.
Invention is credited to Kevin J. Parker, Jose Tamez-Pena, Saara Marjatta Sofia Totterman.
Application Number | 20060122486 11/320743 |
Document ID | / |
Family ID | 24155815 |
Filed Date | 2006-06-08 |
United States Patent
Application |
20060122486 |
Kind Code |
A1 |
Tamez-Pena; Jose ; et
al. |
June 8, 2006 |
Magnetic resonance imaging with resolution and contrast
enhancement
Abstract
MRI scans typically have higher resolution within a slice than
between slices. To improve the resolution, two MRI scans are taken
in different, preferably orthogonal, directions. The scans are
registered by maximizing a correlation between their gradients and
then fused to form a high-resolution image. Multiple receiving
coils can be used. When the images are multispectral, the number of
spectral bands is reduced by transformation of the spectral bands
in order of image contrast and using the transformed spectral bands
with the highest contrast.
Inventors: |
Tamez-Pena; Jose;
(Rochester, NY) ; Totterman; Saara Marjatta Sofia;
(Rochester, NY) ; Parker; Kevin J.; (Rochester,
NY) |
Correspondence
Address: |
BLANK ROME LLP
600 NEW HAMPSHIRE AVENUE, N.W.
WASHINGTON
DC
20037
US
|
Family ID: |
24155815 |
Appl. No.: |
11/320743 |
Filed: |
December 30, 2005 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11110717 |
Apr 21, 2005 |
6984981 |
|
|
11320743 |
Dec 30, 2005 |
|
|
|
09540524 |
Mar 31, 2000 |
6998841 |
|
|
11110717 |
Apr 21, 2005 |
|
|
|
Current U.S.
Class: |
600/410 |
Current CPC
Class: |
G01R 33/4835 20130101;
A61B 5/055 20130101; A61B 5/4528 20130101; G01R 33/5608
20130101 |
Class at
Publication: |
600/410 |
International
Class: |
A61B 5/05 20060101
A61B005/05 |
Claims
1-20. (canceled)
21. A method of forming an image of a subject, the method
comprising: (a) performing an MRI scan on the subject to take image
data having a plurality of spectral bands; and (b) forming the
image from the image data.
22. The method of claim 21, wherein step (b) comprises selecting a
subplurality of the plurality of spectral bands to form the
image.
23. The method of claim 22, wherein the subplurality of the
plurality of spectral bands is selected by ranking the plurality of
spectral bands in order of image contrast and selecting the
spectral bands whose image contrast is highest.
24. The method of claim 23, wherein the plurality of spectral bands
is ranked in order of image contrast by: deriving a covariance
matrix from the plurality of spectral bands; deriving a set of
orthogonal eigenvectors and a corresponding set of eigenvalues from
the covariance matrix; and ranking the orthogonal eigenvectors in
order of their corresponding eigenvalues.
25. The method of claim 21, wherein step (a) is performed with a
plurality of receiving coils.
26. The method of claim 25, wherein step (a) is performed using at
least three receiving coils.
27. The method of claim 26, wherein step (a) is performed using at
least four receiving coils.
28-47. (canceled)
48. A system for forming an image of a subject, the system
comprising: scanning means for performing an MRI scan on the
subject to take image data having a plurality of spectral bands;
and computing means for forming the image from the image data.
49. The system of claim 48, wherein the computing means selects a
subplurality of the plurality of spectral bands to form the
image.
50. The system of claim 49, wherein the subplurality of the
plurality of spectral bands is selected by ranking the plurality of
spectral bands in order of image contrast and selecting the
spectral bands whose image contrast is highest.
51. The system of claim 50, wherein the plurality of spectral bands
is ranked in order of image contrast by: deriving a covariance
matrix from the plurality of spectral bands; deriving a set of
orthogonal eigenvectors and a corresponding set of eigenvalues from
the covariance matrix; and ranking the orthogonal eigenvectors in
order of their corresponding eigenvalues.
52. The system of claim 48, wherein the scanning means comprises a
plurality of receiving coils.
53. The system of claim 52, wherein the plurality of receiving
coils comprises at least three receiving coils.
54. The system of claim 53, wherein the plurality of receiving
coils comprises at least four receiving coils.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] The present invention is directed to a system and method for
improving the resolution and tissue contrast in MRI.
[0003] 2. Description of Related Art
[0004] The best current source of raw image data for observation of
a complex soft tissue and bone structure is magnetic resonance
imaging (MRI). MRI involves the transmission of RF signals of
predetermined frequency (e.g., approximately 15 MHZ in some
machines, the frequency depending upon the magnitude of magnetic
fields employed and the magnetogyric ratio of the atoms to be
imaged). Typically, exciting pulses of RF energy of a specific
frequency are transmitted via an RF coil structure into an object
to be imaged. A short time later, radio-frequency NMR responses are
received via the same or a similar RF coil structure. Imaging
information is derived from such RF responses.
[0005] In MRI, a common imaging technique is the formation of
images of selected planes, or slices, of the subject being imaged.
Typically the subject is located in the static magnetic field with
the physical region of the slice at the geometric center of the
gradient field. Generally, each gradient will exhibit an increasing
field strength on one side of the field center, and a decreasing
field strength on the other side, both variations progressing in
the direction of the particular gradient. The field strength at the
field center will thus correspond to a nominal Larmor frequency for
the MRI system, usually equal to that of the static magnetic field.
The specific component of a gradient which causes the desired slice
to be excited is called the slice selection gradient. Multiple
slices are taken by adjusting the slice selection gradient.
[0006] However, MRI often introduces the following technical
challenges. Many of the anatomical structures to be visualized
require high resolution and present low contrast, since, for
example, many of the musculoskeletal structures to be imaged are
small and intricate. MRI involves the use of local field coils to
generate an electromagnetic field; such local field coils form a
non-uniform illumination field. MRI images can also be noisy.
[0007] In particular, MRI has the following limitations in
resolution and tissue contrast. Although current MRI machines can
achieve relatively high intra-plane resolution, the inter-slice
resolution is not so good as the intra-plane resolution; also, the
inter-slice resolution is limited by the ability of the system to
stimulate a single spatial slice or section. Although tissue
contrast can be adjusted by selecting the right pulse sequence,
analysis of a single pulse sequence is not enough to differentiate
among adjacent similar tissues. In other words, the resolution is
typically poor in the out-of-plane dimension, and the contrast is
typically low between soft tissue structures.
SUMMARY OF THE INVENTION
[0008] It will be readily apparent from the foregoing that a need
exists in the art to overcome the above-noted limitations of
conventional MRI.
[0009] Therefore, it is an object of the present invention to
increase inter-slice resolution.
[0010] It is another object of the present invention to improve
tissue contrast.
[0011] It is still another object of the present invention to
improve inter-slice resolution and tissue contrast
simultaneously.
[0012] It is yet another object to provide a simple technique for
image registration.
[0013] To achieve the above and other objects, the present
invention is directed to a system and method for creating
high-resolution MRI volumes and also high-resolution,
multi-spectral MRI volumes. At least one additional scan is
obtained in an orthogonal direction. Then, through a data fusion
technique, the information from an original scan and an orthogonal
scan are combined, so as to produce a high-resolution, 3D volume.
In addition, one may use a different pulse sequence in the original
orientation or in an orthogonal orientation, and a data fusion
technique can be applied to register the information and then
visualize a high-resolution, multi-spectral volume.
BRIEF DESCRIPTION OF THE DRAWINGS
[0014] A preferred embodiment will be set forth in detail with
reference to the drawings, in which:
[0015] FIG. 1 shows a block diagram of an MRI system according to
the preferred embodiment;
[0016] FIGS. 2 and 3 show flow charts showing steps performed in
registering and fusing two images;
[0017] FIG. 4 shows a blurring effect caused by the correlation of
the two images;
[0018] FIGS. 5A-5C show a typical signal, its gradient and a
comparison of the autocorrelations of the signal and its gradient,
respectively;
[0019] FIGS. 6A and 6B show two voxels scanned in orthogonal
directions;
[0020] FIG. 6C shows the problem of deriving high-resolution
information from the voxels of FIGS. 6A and 6B;
[0021] FIGS. 7A-7I show comparisons between the individual scans
and the fused image;
[0022] FIGS. 8A-8I show a comparison among simple fusion without
registration, simple fusion after registration and complete
fusion;
[0023] FIGS. 9A-9I show fusion of orthogonal images without
correlation;
[0024] FIGS. 10A and 10B show images taken with three and four
local receiver coils, respectively;
[0025] FIGS. 11A and 11B show a two-band spectral image of a knee;
and
[0026] FIGS. 11C and 11D show principal components of the image of
FIGS. 11A and 11B.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0027] A preferred embodiment of the present invention will now be
set forth in detail with reference to the drawings.
[0028] FIG. 1 shows a block diagram of an MRI system 100 on which
the present invention can be implemented. The system 100 uses an RF
coil 102 and a gradient coil 104 to apply the required RF and
gradient fields to the subject S. A spectrometer 106, acting under
the control of a computer 108, generates gradient signals which are
amplified by an X amplifier 110, a Y amplifier 112 and a Z
amplifier 114 and applied to the gradient coil 104 to produce the
gradient fields. The spectrometer 106 also generates RF signals
which are amplified by an RF amplifier 116 and applied to the RF
coil 102 to produce the RF fields. The free induction decay
radiation from the sample S is detected by the RF coil 102 or by
one or more local receiving coils 118 and applied to the
spectrometer 106, where it is converted into a signal which the
computer 108 can analyze.
[0029] The computer 108 should be sufficiently powerful to run a
mathematical analysis package such as AVS, a product of Advanced
Visualization Systems of Waltham, Mass., U.S.A. Examples are the
Apple Power Macintosh and any IBM-compatible microcomputer capable
of running Windows 95, 98 or NT. The significance of the local
receiving coils 118, and particularly of the number used, will be
explained in detail below. The other components of the system 100
will be familiar to those skilled in the art and will therefore not
be described in detail here.
[0030] The various techniques to enhance the images will now be
described in detail.
[0031] Inter-Slice Resolution
[0032] The inter-slice resolution problem is solved by using two
volumetric data sets where scanning directions are orthogonal to
each other, and fusing them in a single high-resolution image. FIG.
2 shows an overview of the process. First, in step 202, a first
scan of the subject is taken. In step 204, a second scan of the
subject is taken in a direction orthogonal to that of the first
scan. Third, in step 206, the images are registered. Finally, in
step 208, the images are fused. FIG. 3 shows the steps involved in
image registration. In step 302, gradients of the image data are
formed. In step 304, the gradients are correlated. In step 306, the
correlation is maximized through a hill-climbing technique.
[0033] Although the fusion of the two volumes seems to be trivial,
two issues have to be taken into account: 1) the registration among
volumes scanned at different time (step 206) and 2) the overlapping
of sampling voxel volumes (step 208). Those issues are handled in
ways which will now be described.
[0034] Image Registration (Step 206)
[0035] The goal of image registration is to create a
high-resolution 3D image from the fusion of the two data sets.
Therefore, the registration of both volumes has to be as accurate
as the in-plane resolution. The preferred embodiment provides a
very simple technique to register two very similar orthogonal MRI
images. The registration is done by assuming a simple translation
model and neglecting the rotation among the two volumes, thus,
providing a fair model for small and involuntary human motion
between scans.
[0036] An unsupervised registration algorithm finds the point (x,
y, z) where the correlation between the two data sets is maximum.
The Schwartz inequality identifies that point as the point where
they match. Given two functions u(x, y, z) and v(x, y, z), the
correlation is given by: r .function. ( x , y , z ) = u .function.
( x , y , z ) * v .function. ( x , y , z ) = .intg. .intg. .intg. u
.function. ( .alpha. , .beta. , .gamma. ) .times. v .function. ( x
+ .alpha. , y + .beta. , z + .gamma. ) .times. d .alpha. .times. d
.beta. .times. d .gamma. . ##EQU1## If u(x, y, z) is just a
displaced version of v(x, y, z), or in other words, u(x, y,
z)=v(x+.DELTA.x, y+.DELTA.y, z+.DELTA.z), then the maximum is at
(-.DELTA.x, -.DELTA.y, -.DELTA.z), the displacement between the
functions.
[0037] In MRI, every voxel in the magnetic resonance image observes
the average magnetization of an ensemble of protons in a small
volume. The voxel can therefore be modeled as the correlation of
the continuous image by a small 3D window. Let g.sub.1(x, y, z) and
g.sub.2(x, y, z) be the sampled volumes at two orthogonal
directions, which are given by:
g.sub.1(x,y,z)=f(x,y,z)*w.sub.1(x,y,z)*.PI.(x,y,z)
g.sub.2(x,y,z)=f(x+.DELTA.x,y+.DELTA.y,z+.DELTA.z)*w.sub.2(x,y,z)*.PI.(x,-
y,z) where w.sub.1(x, y, z) and w.sub.2(x, y, z) are the 3D windows
for the two orthogonal scanning directions, (.DELTA.x, .DELTA.y,
.DELTA.z) is a small displacement, and .PI.(x, y, z) is the
sampling function. Therefore, the correlation of the two sampled
volumes is r .function. ( x , y , z ) = .times. g 1 .function. ( x
, y , z ) * g 2 .function. ( x , y , z ) = .times. f .function. ( x
+ .DELTA. .times. .times. x , y + .DELTA. .times. .times. y , z +
.DELTA. .times. .times. z ) * f .function. ( x , y , z ) * w 1
.function. ( x , y , z ) * .times. w 2 .function. ( x , y , z ) *
.PI. .function. ( x , y , z ) . ##EQU2## That is just a blurred
version of the original correlation; the exact location of the
maxima is shaped by the blurring function h(x, y, z)=w.sub.1(x, y,
z)*w.sub.2(x, y, z). That is, the correlation is distorted by the
function h(x, y, z). FIG. 4 shows an idealized window function for
w.sub.1(x, y, z) and w.sub.2(x, y, z) and its corresponding
supporting region of the blurring function, h(x, y, z).
[0038] Finding the displacement using the above procedure works
fine for noise-free data, but noise makes the search more
difficult. Let g.sub.1(x, y, z)=(f(x, y, z)+n.sub.1(x, y,
z))*w.sub.1(x, y, z) and g.sub.2(x, y, z)=(x+.DELTA.x, y+.DELTA.y,
z+.DELTA.z)+n.sub.2(x, y, z))*w.sub.2(x, y, z) be the corresponding
noisy volumes, where n.sub.1(x, y, z) and n.sub.2(x, y, z) are two
uncorrelated noise sources. Thus, the correlation of g.sub.1 and
g.sub.2 is given by r .function. ( x , y , z ) = .times. g 1
.function. ( x , y , z ) * g 2 .function. ( x , y , z ) = .times. [
f .function. ( x + .DELTA. .times. .times. x , y + .DELTA. .times.
.times. y , z + .DELTA. .times. .times. z ) * f .function. ( x , y
, z ) + .times. f .function. ( x + .DELTA. .times. .times. x , y +
.DELTA. .times. .times. y , z + .DELTA. .times. .times. z ) * n 1
.function. ( x , y , z ) + f .function. ( x , y , z ) * .times. n 2
.function. ( x , y , z ) ] * h .function. ( x , y , z ) * .PI.
.function. ( x , y , z ) , ##EQU3## and the maximum is no longer
guaranteed to be given by the displacement, especially for
functions with smooth autocorrelation functions like standard MRI.
The smooth autocorrelation functions make the error a function of
the noise power. Autocorrelation functions whose shapes are closer
to a Dirac delta function .delta.(x, y, z) are less sensitive to
noise, which is the reason why many registration algorithms work
with edges. The autocorrelation function of the gradient magnitude
of standard MRI is closer to a Dirac delta function. Therefore, the
registration of the gradient is less sensitive to noise.
[0039] FIGS. 5A-5C shows a 1D example of the effect of the
derivative on the autocorrelation function of a band-limited
signal. FIG. 5A shows an original signal f(x). FIG. 5B shows a
smooth estimate of the magnitude of the derivative, namely,
g(x)=|f(x)*[-1 -1 0 0 0 1 1]| FIG. 5C shows autocorrelations; the
dashed curve represents f*f, while the curve shown in crosses
represents g*g. For that example, a smooth derivative operator is
used to reduce the noise level.
[0040] For the above reasons, the automatic registration is based
on finding the maximum on the correlation among the magnitude
gradient of the two magnetic resonance images: ( .DELTA. .times.
.times. x , .DELTA. .times. .times. y , .DELTA. .times. .times. z )
= Arg .times. max ( .DELTA. .times. .times. x , .DELTA. .times.
.times. y , .DELTA. .times. .times. z ) .times. .intg. .intg.
.intg. .gradient. g 1 .function. ( .alpha. , .beta. , .gamma. )
.times. .gradient. g 2 .function. ( .DELTA. .times. .times. x +
.alpha. , .DELTA. .times. .times. y + .beta. , .DELTA. .times.
.times. z + .gamma. ) .times. d .alpha. .times. d .beta. .times. d
.gamma. . ##EQU4##
[0041] In sampled images the gradient .gradient. can be
approximated by finite differences: .gradient. g .function. ( x , y
, z ) = d x .function. ( x , y , z ) 2 + d y .function. ( x , y , z
) 2 + d z .function. ( x , y , z ) 2 , .times. where ##EQU5## d x
.function. ( x , y , z ) = l .function. ( x , y , z ) * g
.function. ( x + .delta. .times. .times. x , y , z ) - l .function.
( x , y , z ) * g .function. ( x - .delta. .times. .times. x , y ,
z ) 2 .times. .delta. .times. .times. x , .times. d y .function. (
x , y , z ) = l .function. ( x , y , z ) * g .function. ( x , y +
.delta. .times. .times. y , z ) - l .function. ( x , y , z ) * g
.function. ( x , y - .delta. .times. .times. y , z ) 2 .times.
.delta. .times. .times. y , .times. d z .function. ( x , y , z ) =
l .function. ( x , y , z ) * g .function. ( x , y , z + .delta.
.times. .times. z ) - l .function. ( x , y , z ) * g .function. ( x
, y , z - .delta. .times. .times. z ) 2 .times. .delta. .times.
.times. z , ##EQU5.2## where .delta.x, .delta.y, .delta.z are the
sampling rates, and l(x, y, z) is a low pass filter used to remove
noise from the images and to compensate the differences between
in-slice sampling and inter-slice sampling.
[0042] The maximization can be done using any standard maximization
technique. The preferred embodiment uses a simple hill-climbing
technique because of the small displacements. The hill-climbing
technique evaluates the correlation at the six orthogonal
directions: up, down, left, right, front and back. The direction
that has the biggest value is chosen as the next position. That
simple technique works well for the registration of two orthogonal
data sets, as the one expects for involuntary motion during
scans.
[0043] To avoid being trapped in local maxima and to speed up the
process, a multi-resolution approach can be used. That
multi-resolution approach selects the hill-climbing step as half
the size of the previous step. Five different resolutions are used.
The coarsest resolution selected is twice the in-plane resolution
of the system, and the smallest size is just 25 percent of the
in-plane resolution.
[0044] Even with a very simple optimization approach, the
computation of the correlation of the whole data set can be time
consuming; therefore, the registration of two images can take time.
To speed up the correlation of the two data sets, just a small
subsample of the points with very high gradient are selected for
use in the correlation process.
[0045] Some images suffer from a small rotation. In that case, the
algorithm is extended to search for the image rotation. The same
hill-climbing technique is used to find the rotation between
images; but instead of doing the search in a three-dimensional
space, the algorithm has to look at a six-dimensional space. That
search space includes the three displacements and three rotations
along each axis. At each step the rotation matrix is updated and
used to compensate for the small rotation between images.
[0046] Image Fusion (Step 208)
[0047] Once the two images are registered, an isotropic high
resolution image is created from them. Due to the different shape
between voxel sampling volumes (w.sub.2(x, y, z) and w.sub.1(x, y,
z)), one has to be careful when estimating every high resolution
voxel value from the input data. Assume that the first image has
been scanned in the x-direction and the second has been scanned in
the y-direction. Therefore, there is high-resolution information in
the z-direction in both images.
[0048] FIGS. 6A and 6B show the voxel shapes of the two input
images, where the in-slice resolution is equal, and the inter-slice
resolution is four times lower. Given that configuration, the
problem of filling the high-resolution volume is a 2D problem. In a
single 4.times.4-voxel window of the high resolution image, as seen
in FIG. 6C, then for every 16 high-resolution voxels there are only
8 known low-resolution voxels; therefore, that is an ill posed
problem.
[0049] To address that problem, assume that every high-resolution
voxel is just a linear combination of the two low-resolution
functions: g(x,y,z)=h.sub.1(x/s.sub.d,y,z)+h.sub.2(x,y/s.sub.d,z),
where h.sub.1, h.sub.2 are two functions which are back projected
in such a way that g 1 .function. ( x / s d , y , z ) =
.A-inverted. x .di-elect cons. w 1 .times. g .function. ( x , y , z
) , .times. g 2 .function. ( x , y / s d , z ) = .A-inverted. y
.di-elect cons. w 2 .times. g .function. ( x , y , z ) . ##EQU6##
That represents a linear system with the same number of knowns as
unknowns. The known values are the observed image voxels, while the
values to back-project which match the observation are estimated.
Noise and inhomogeneous sampling make the problem a little bit
harder; but that linear system can efficiently be solved using
projection on convex sets (POCS). Although, in theory, all the
components have to be orthogonally projected, it can be shown that
the following projecting scheme also works: h 1 k + 1 .function. (
x / s d , y , z ) = h 1 k .function. ( x / s d , y , z ) - .alpha.
.times. ( nh 1 k .function. ( x / s d , y , z ) + .A-inverted. y
.di-elect cons. w 1 .times. h 2 k .function. ( x , , y / s d , z )
- g 1 .function. ( x / s d , y , z ) ) .times. n n 2 + m , .times.
h 2 k + 1 .function. ( x , y / s d , z ) = h 2 k .function. ( x , y
/ s d , z ) - .alpha. .times. ( mh 2 k .function. ( x , y / s d , z
) + .A-inverted. y .di-elect cons. w 2 .times. h 1 k .function. ( x
/ s d , y , z ) - g 2 .function. ( x , y / s d , z ) ) .times. m n
+ m 2 , ##EQU7## where 0<.alpha.<1, n=the window size of
w.sub.1, m=the window size of w.sub.2, h.sub.1.sup.0(x, y, z)=g(x,
y, z) and h.sub.2.sup.0(x, y, z)=g.sub.2(x, y, z) are the initial
guesses for the estimation of back-projected functions. The
advantage of that approach over standard orthogonal projection is
that is equations are simpler and that they can be implemented
efficiently on a computer.
[0050] Experimental Data
[0051] Some experimental data produced by the above technique will
now be described.
[0052] FIGS. 7A, 7B and 7C respectively show the original MRI
sagittal scan, the original MRI axial scan and the fused image of a
human shoulder seen along an axial view. FIGS. 7D, 7E and 7F show
the same, except seen along a sagittal view. FIGS. 7G, 7H and 7I
show the same, except seen along a coronal view. In all three
cases, the image is noticeably improved.
[0053] FIGS. 8A, 8B and 8C show axial, sagittal and coronal slices,
respectively, of simple fusion without registration. FIGS. 8D, 8E
and 8F show the same slices with simple fusion after registration.
FIGS. 8G, 8H and 8I show the same slices with complete image
fusion. The simple fusion is g(x, y, z)=0.5 g.sub.1(x, y, z)+0.5
g.sub.2(x, y, z), and the complete fusion is g(x, y, z)=h.sub.1(x,
y, z)+h.sub.2(x, y, z), wherein h.sub.1 and h.sub.2 are the two
functions which minimize the reconstruction error.
[0054] FIGS. 9A-9I show fusion of orthogonal images without
correlation. FIGS. 9A, 9B and 9C show axial views of the original
MRI sagittal scan, the original axial scan and the fused image,
respectively, for an axial view. FIGS. 9D, 9E and 9F show the same
for a sagittal view. FIGS. 9G, 9H and 9I show the same for a
coronal view.
[0055] Multiple Local Coil Receivers and Multispectral Imaging
[0056] Good signal-to-noise ratio is very important for an
unsupervised segmentation algorithm. Even more important is the
contrast-to-noise ratio among neighboring tissues. When local
receiving coils are used, the signal from points far from the coil
location is weak; therefore, contrast among tissues located far
from the receiving coil is low. Some researchers have proposed
several software alternatives to correct this signal fading, but
this will increase the noise levels as well. Thus, it will not
solve the problem. The preferred embodiment uses two or more
receiving coils, which will improve the signal reception at far
locations.
[0057] FIGS. 10A and 10B show the advantage of using multiple
coils. FIG. 10A shows an MRI image of a knee using three surface
coils. FIG. 10B shows an MRI image of the same knee using four
surface coils.
[0058] Multispectral images will now be considered. Such images can
be analogized to multispectral optical images, in which, for
example, red, blue and green images are combined to create a single
color image.
[0059] FIGS. 11A and 11B show a two-band spectral image of a knee.
FIG. 11A shows a cross section of a fat suppression MRI scan of the
knee, where fat, and bone tissues have almost the same low density,
cartilage has a very high density, and muscle tissue has a medium
density. FIG. 11B shows the same knee, but now, muscle tissue and
cartilage have the same density, making them very hard to
differentiate. Those images clearly show the advantage of the
multispectral image approach in describing the anatomy.
[0060] The analysis of a multispectral image is more complex than
that of a single-spectrum image. One way to simplify the analysis
is to reduce the number of bands by transforming an N-band
multispectral image in such a way that passes the most relevant
image into an (N-n)-band image. The transform that minimizes the
square error between the (N-n)-band image and the N-band image is
the discrete Karhuen-Loeve (K-L) transform. The resulting
individual images from the transformed spectral image after
applying the K-L transform are typically known as the principal
components of the image.
[0061] Let the voxel x be an N-dimensional vector whose elements
are the voxel density from each individual pulse sequence. Then the
vector mean value of the image is defined as m x = E .function. [ x
] = 1 M .times. k = 1 M .times. x k , ##EQU8## and the covariance
matrix is defined as C x = E .function. [ ( x - m x ) .times. ( x -
m x ) 2 ] = 1 M .times. k = 1 M .times. x k .times. x k T - m k
.times. m k T , ##EQU9## where M is the number of voxels in the
image.
[0062] Because C.sub.x is real and symmetric, a set of n orthogonal
eigenvectors can be found. Let e.sub.i and .lamda..sub.i, i=1, 2, .
. . N, be the eigenvectors and the corresponding eigenvalues of
C.sub.x, so that .lamda..sub.j>.lamda..sub.j+1. Let A be the
matrix formed with the eigenvectors of C.sub.x. Then the
transformation y=A(x-m).sub.x is the discrete K-L transform, and
y.sub.i, i=1, 2, . . . N, are the components of a multispectral
image. FIGS. 11C and 11D show the principal components of the
two-band spectral image shown in FIGS. 11A and 11B. The advantage
of the K-L decomposition of a multispectral image is that the image
with the highest contrast is associated with the highest eigenvalue
of the correlation matrix, and the image associated with the
smallest eigenvalues usually is irrelevant.
[0063] The high-contrast, multispectral images thus formed can be
utilized for diagnosis and for input to post-processing systems,
such as three-dimensional rendering and visualization systems. If
the multispectral data are from orthogonal planes or are acquired
with some misregistration, the registration and orthogonal fusion
steps described herein can be employed to enhance the resolution
and contrast.
[0064] While a preferred embodiment of the present invention has
been set forth above, those skilled in the art who have reviewed
the present disclosure will readily appreciate that other
embodiments can be realized within the scope of the present
invention. For example, while the invention has been disclosed as
used with the hardware of FIG. 1, other suitable MRI hardware can
be used. For that matter, the invention can be adapted to imaging
techniques other than MRI, such as tomography. Also, while the
scans are disclosed as being in orthogonal directions, they can be
taken in two different but non-orthogonal directions. Therefore,
the present invention should be construed as limited only by the
appended claims.
* * * * *