Image analysis – Image enhancement or restoration – Object boundary expansion or contraction
Reexamination Certificate
1999-06-14
2003-02-18
Rogers, Scott (Department: 2624)
Image analysis
Image enhancement or restoration
Object boundary expansion or contraction
C382S275000, C382S308000, C382S128000, C382S154000, C600S410000
Reexamination Certificate
active
06522786
ABSTRACT:
BACKGROUND OF THE INVENTION
The field of this invention is nuclear magnetic resonance imaging methods and systems. More particularly, the invention relates to a method and apparatus for reducing noise in a three dimensional data array generated using magnetic resonance imaging techniques.
Any nucleus which possesses a magnetic moment attempts to align itself with the direction of the magnetic field in which it is located. In doing so, however, the nucleus precesses around this direction at a characteristic angular frequency (Larmor frequency) which is dependent on the strength of the magnetic field and on the properties of the specific nuclear species (the magnetogyric constant &ggr; of the nucleus). Nuclei which exhibit this phenomena are referred to herein as “spins”.
While many different tissue samples and various bodies may be examined using NMR imaging, in order to further simplify this explanation the invention is described in the context of an exemplary transaxial volume through a patient's body wherein the volume includes the patient's heart and the volume will be referred to as a region of interest. In addition, it will be assumed that an NMR imaging system includes a three dimensional imaging area having Cartesian coordinate x, y and z axes and that the patient is positioned within the imaging area with the patient's height (i.e. from head to feet) defining an axis along the z axis.
When the region of interest is subjected to a uniform magnetic field (polarizing field B
0
), the individual magnetic moments of the nuclear spins in the region attempt to align with the polarizing field, but precess about the direction of the field in random order at their characteristic angular or Larmor frequencies, producing a net magnetic moment M
z
in the direction of the polarizing field.
If the region of interest is subjected to a magnetic field (excitation field B
1
) which is in the x-y plane and which is near the Larmor frequency, the net aligned moment M
z
may be “tipped” into the x-y plane to produce a net transverse magnetic moment which is rotating or spinning in the xy plane at the Larmor frequency.
The NMR signal which is emitted by the excited spins after the excitation signal B
1
is terminated is a function of physical properties of the spin which generates the signal. These emitted NMR signals are digitized and processed to generate an NMR data set.
To determine the point of origin of an NMR signal, each NMR signal is encoded with spatial information, such as by the “spin-warp” technique, discussed by W. A. Edelstein et al. in “Spin Warp NMR Imaging and Applications to Human Whole-Body Imaging”,
Physics in Medicine and Biology,
Vol. 25, pp. 751-756 (1980) which is incorporated herein by reference.
According to the spin-warp scheme, spatial encoding is accomplished by employing three magnetic gradient fields (G
x
, G
y
, and G
z
) which have the same direction as polarizing field B
0
and which have gradients along the x, y and z axes, respectively. By controlling the strength of these gradients during each NMR cycle, the spatial distribution of spin excitation can be controlled and the point of origin of the resulting NMR signals can be identified.
A useful acquisition technique is the slice or two dimensional technique wherein NMR data are acquired for a single transaxial slice of a region of interest at one time. The invention is described in the context of slice imaging wherein several slices are acquired consecutively and are “stacked” to form a three dimensional data set.
To determine the z-axis origin of a signal during slice data acquisition, signal generation is limited to a specific transaxial slice of the region of interest using gradient field G
z
. To this end, the Larmor frequency F of a spin can be expressed as:
F
=(
B
0
+B
z
)&ggr; (1)
where B
z
is essentially the strength of gradient G
z
within a specific transaxial slice of the region of interest and is the magnetogyric constant of the nucleus of the material in which the field is generated. Because the gradient field strength varies along the z-axis, each z-axis slice has a different Larmor frequency F. When the excitation signal B
0
is provided at a specific excitation frequency, only spins within the “selected” z-axis slice which are at the excitation frequency are tipped. Therefore, when the excitation signal B
0
is turned off, only spins within the selected z-axis slice generate NMR signals.
To spatially encode NMR signals along the x axis, excitation signal B
0
is provided at a small range of frequencies. The x axis gradient G
x
is small enough that all of the spins along the x axis have Larmor frequencies within the small range of excitation signal frequencies and therefore each of the spins along the x axis generates an NMR signal when the excitation signal is turned off, each x-axis signal having a unique and identifiable frequency. Hence, x-axis position can be determined by identifying the frequency of each NMR signal received during an acquisition. This type of encoding is commonly referred to as frequency encoding.
To encode y axis position within NMR signals, the y axis gradient G
y
is employed to cause spins along the y axis to have different phases; therefore, resulting NMR signals from spins along the y axis have different phases which can be used to determine y axis position. Because y axis position is encoded using signal phase, this type of encoding is commonly referred to as phase encoding.
After data have been acquired for one region of interest slice, the acquisition process is repeated for adjacent region of interest slices until data have been acquired for every slice within the region of interest. After digitizing and processing, the slice data are combined to provide a three dimensional data point (TDDP) array. The TDDP array includes a plurality of data points distributed at regular parallelepiped positions in a three dimensional lattice within the region of interest, at least one value (Vxyz) being characteristic of a physical property of the region of interest associated with each respective one of the lattice positions. Each cubically adjacent set of eight such positions defines a cubic volume referred to hereinafter as a “voxel”, a physical property value being specified for each of the eight voxel vertices.
After a complete TDDP array has been acquired and stored, the array can be used to form an image of the region of interest using one of many well known reconstruction techniques.
For the purposes of this explanation, signals which are generated by spins and are characteristic of the property of the region of interest being detected will be referred to as “true” signals, signal components which are randomly generated within a region of interest will be referred to generally as “noise” and the combination of true signals and noise will be referred to as a “combined” signal.
While extreme measures are taken when designing an NMR system to minimize stray and random magnetic fields and signals within the region of interest during a data acquisition period, noise often occurs in two forms: first, as a background distortion exhibiting a low and relatively constant amplitude throughout a region of interest, and second, with appreciable amplitude caused by localized magnetic fields within the region of interest. The latter type of noise, being localized, will be referred to hereinafter as “localized noise”.
Unfortunately, extremely sensitive sensing coils required to detect low amplitude true signals also detect an appreciable amount of background noise from within the region of interest during an acquisition period. Therefore, after a data acquisition period, each TDDP array data point typically includes both a true signal component and a noise component (i.e each data point value constitutes a combined signal). In addition, some data points are dominated by a localized noise component.
While an image can be generated using combined signals, the noise components reduce image clarity and minimize diagnostic usefulness of the
Patnode Patrick K.
Rogers Scott
Testa Jean K.
LandOfFree
Gradient filter and method for three dimensional NMR data set does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Gradient filter and method for three dimensional NMR data set, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Gradient filter and method for three dimensional NMR data set will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3121310