Pulse or digital communications – Repeaters – Testing
Reexamination Certificate
1999-08-16
2001-07-17
Bruce, David V. (Department: 2882)
Pulse or digital communications
Repeaters
Testing
C378S015000, C378S901000
Reexamination Certificate
active
06263008
ABSTRACT:
BACKGROUND OF THE INVENTION
In an X-ray computed tomography (CT) system, an energy source irradiates X-ray beams through an object, and a detector array senses and measures the intensity of the attenuated X-ray beams across a thin section of the object. The X-ray beam intensity level incident at each detector is digitized and converted to a value representing the line integral, referred to in the art as a “projection”, of the object along the X-ray beam path.
In a stationary scan configuration, or “non-helical” scan mode, the object is fixed in position during each scan, while in a translational scan, or “helical” scan mode, the object translates in the same direction as the axis of rotation (z-axis) during a scan, improving system throughput.
For third-generation CT systems, during a scan, an X-ray source and a detector array are mounted on a gantry and rotate together about an object. Successive sets of projections of the object are recorded at incremental gantry rotation angles. At each rotation angle, the collected projections represent a projection profile of the object at that angle. With a set of projection profiles over many view angles, an image of the object across the scanned section, or “slice”, can be generated in a process known as reconstruction, which involves a convolution and back projection of the collected projections.
In a “conventional” CT system, the detector array comprises a single-row detector array, while a “cone-beam” CT system employs a two-dimensional detector array typically having multiple rows and multiple columns of detectors. A cone-beam CT system allows for multiple slices of an object to be scanned simultaneously. An example of a non-helical cone-beam scan technique is described in Feldkamp et al., “Practical Cone-Beam Algorithm”, J. Opt. Soc. Am. A/Vol 1, p612, No. 6, June 1984. An example of a helical cone-beam scan is described in U.S. Pat. No. 5,430,783, issued Jul. 4, 1995 to Hu, et al.
With reference to
FIG. 1
, which is a schematic diagram of a cone-beam system, if the height (the length along the z-axis) of the ith detector row is &Dgr;H
i
, then the equivalent scaled height &Dgr;h
i
of the detector row at isocenter
14
(in this case, assume the isocenter to be along the z-axis) is:
&Dgr;
h
i
=&Dgr;H
i
R/D
(1)
where R and D represent distances from the X-ray focal spot
10
to the detector array
12
and from the x-ray focal spot
10
to the system isocenter
14
, respectively; and the ratio R/D represents scaling factor.
In an “equal height” detector system for a cone-beam scanner, row heights &Dgr;H
i
are the same for all rows of detectors 1 . . . M. In a “non-equal height” detector system, the rows may be of different heights. In one such non-equal height system, the individual row heights &Dgr;H
i
are configured according to certain integer multiples such that adjacent rows can be combined to provide the equivalent of an effective constant group height &Dgr;H. For example, assuming a multiple-row detector array having 8 individual rows of respective detector heights: 5t, 2t, 2t, 1t, 1t, 2t, 2t, 5t, as shown in
FIG. 5
, the data collected at particular row groupings of the array can be combined to become 4 rows of detectors at a constant height of &Dgr;H 5t (i.e., combined to 5t, 2t+2t+1t, 1t+2t+2t, 5t, respectively). A number of additional combinations can also be realized. This type of detector configuration is disclosed in U.S. patent application Ser. No. 09/159,067, filed Sep. 23, 1998, commonly owned with the present application, and incorporated herein by reference.
In a non-helical scan cone-beam CT system, the resolution of the reconstructed image along the z-axis, which in turn is referred to herein as the “slice width” or “slice thickness”, is determined by the detector height at isocenter
14
, &Dgr;h
i
. This is analogous to the slice thickness of a conventional CT system having a single row of detectors, except that in the cone-beam system multiple slices are scanned simultaneously by the multiple rows. In the preceding example with four groups of rows at a constant group height of detection of &Dgr;H=5t, all four slices will exhibit the same slice thickness of &Dgr;h=5t R/D.
However, for a helical scan, effective slice thickness broadens considerably. This is due to the fact that data for each slice is acquired by different rows or groups of rows of detectors during the course of the helical scan. At each view angle, data for each slice is, in general, acquired by either one or two rows or groups of rows of detectors due to translation of the object during a scan, and effective height of detection is thus varied between one and two rows or groups of rows of detectors. Consequently, the slice thickness of a helical scan is considerably broadened, as compared to a non-helical scan. Furthermore, the effective height of detection is inconsistent, in that it varies with view angle. Reconstruction artifacts are thus introduced due to variation in the effective height of detection at the multiple view angles.
In this manner, the slice profile is not sharply defined due to broadening of the detection, and the reconstructed image is prone to contain artifacts due to inconsistency of the projection data. To reduce the broadening of the slice, it has been suggested to first reconstruct multiple thin slices by dividing a row of detectors into a set of multiple sub-rows of detectors, and to independently reconstruct a plurality of sub-slices using the set of multiple sub-rows. This is followed by combining the sub-slices into a single composite slice. Such a technique results in a composite slice having a slice thickness considerably greater than each of the individual sub-slices, but having substantially improved slice profile as described in U.S. Pat. No. 5,430,783, issued Jul. 4, 1995 to Hu et al., incorporated herein by reference. However, this approach is computationally expensive in that it requires multiple convolutions and back projections for each sub-slice.
SUMMARY OF THE INVENTION
To address the shortcomings of the prior techniques, the present invention provides a method of and apparatus for mitigating data inconsistency leading to the reconstruction artifacts described above, and for providing a sharper slice profile having improved precision. An over-sampling detector array and improved re-sampling technique are provided to prepare projection data for reconstruction. The over-sampling detector array employs a plurality of rows of detectors, with the height of each detector row being less than the slice thickness. The improved re-sampling technique involves a re-sampling of the over-sampled data, and combination thereof prior to reconstruction, thereby enhancing resulting image quality in a manner that improves system throughput.
In a first aspect, the present invention is directed to a method of reconstructing an object image as successive slices of a predetermined slice thickness in a helical cone-beam computed tomography scanner. Projection data are over-sampled with a two-dimensional detector array arranged in rows and columns. The detector array includes detector rows each having a height less than the slice thickness. The projection data are re-sampled relative to a system translation axis according to the predetermined slice thickness. The re-sampled projection data are next reconstructed to generate an object image.
Over-sampling may comprise over-sampling projection data with a detector array having row heights that are equal (an “equal-height” detector array) or unequal (a “non-equal-height” detector array). Projection data re-sampling preferably comprises identifying the boundaries of each slice along the translation axis and summing the data values for detector rows between the boundaries. Boundary detector rows containing the boundaries are identified and the projection data collected by the boundary detector rows are weighted according to the relative position of the boundary with respect to the detector row along the translation axis. The lo
Analogic Corporation
Bruce David V.
McDermott & Will & Emery
LandOfFree
Over-sampling detector array and re-sampling technique for a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Over-sampling detector array and re-sampling technique for a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Over-sampling detector array and re-sampling technique for a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2512133