X-ray or gamma ray systems or devices – Specific application – Tomography
Reexamination Certificate
1998-08-21
2002-10-01
Bruce, David V. (Department: 2882)
X-ray or gamma ray systems or devices
Specific application
Tomography
C250S311000
Reexamination Certificate
active
06459758
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates to the field of tomography and, in particular, to the reconstruction of images from emission tomography of microscopic samples, for example, as applied to the internal structure of semiconductors. A non-linear programming algorithm, known as the expectation/maximization (EM) algorithm, is bounded and applied to the solution of the data recovered by discrete emission tomography to reproduce an image for display.
2. Description of the Relevant Art
U.S. Pat. No. 5,659,175, to Fishburn et al. describes apparatus and a method for tomography of microscopic samples. For example, let S be a finite subset of a lattice Z and let
be a finite collection of subsets of Z. For each L&egr;
, let v(L) be the number of points in L∩S. Fishburn et al. recognized that the integer programming problem could be translated into a problem involving fuzzy sets that can be solved using linear programming. The problem of reconstructing the set S from knowledge of {v(L); L&egr;
} when
is a collection of lines in Z in the main directions (the subject of discrete tomography and lines in main directions will be further described herein) is also considered. The approach in the '175 patent is based on (a) relaxing the constraint that S be a set to requiring that S be a fuzzy set, and (b) using a linear program to find a feasible fuzzy set. From a practical point of view (a) amounts to replacing
v
(
L
)
=
∑
z
∈
Z
X
S
(
z
)
X
L
(
z
)
,
L
∈
ℒ
(1.1)
where X
A
denotes the indicator function of a set A, with
v
(
L
)
=
∑
z
∈
Z
f
(
z
)
X
L
(
z
)
,
L
∈
ℒ
,
0
≤
f
(
z
)
≤
1
,
z
∈
Z
(1.2)
and (b) amounts to finding a feasible solution in f(z), z&egr;Z, for the constrained linear system (ii), using linear programming. The approach has important scientific applications in high-resolution electron-microscopy of crystals.
The problem of reconstructing an image for display from probabilities of occupancy of individual lattice sites within a crystal is described by the '175 patent in some detail, but some of the discussion is repeated here with reference to
FIGS. 7-14
which correspond to
FIGS. 1-7
and
13
of the '175 patent. In tomographic radiography, projection radiography is done to capture the amount of attenuation of a beam as it passes through an object as a black level but repeated many times. The repetitions produce a collection of data which is used to generate an image of the cross section of an object. That is, multiple “side views” of the object are used to generate a cross sectional view called a phantom.
Referring first to
FIG. 7
, some of the basic principles of tomography will be briefly described. The object “3” under scan of
FIG. 7A
is conceptually divided into pixels P, as in FIG.
7
B. Each ray R of radiation passing through the body of object “3” is attenuated by the sequence of pixels through which it passes according to the attenuation coefficient of each pixel, unknown in real-life except when the imaged object is an experimental “phantom”. A collection of parallel rays, spanning from point C to point D, produces one of the projection images. The amount of attenuation of each ray is measured using a detector for producing a numerical output. Then, an equation is derived for each ray. A simplified example is given by FIG.
7
C.
If the initial intensity of the ray is I
o
, and if the intensity of the ray after passing through the object is I
f
then Equation (1) is obtained for FIG.
7
C:
I
f
=[I
o
e
−&agr;
1
L
1
][e
−&agr;
2
L
2
][e
−&agr;
3
L
3
] (1)
Each bracketed term in equation (1) corresponds to a position indicated in FIG.
7
C.
Equation (1) results from a single ray. Since multiple rays produce each image, multiple equations are obtained for each image. Further, since numerous images are obtained, numerous sets of equations, each containing multiple equations, are obtained.
The numerous sets, of multiple equations each, are solved for the attenuation variables &agr;, three of which are found in Equation (1). Each &agr; represents the attenuation coefficient of a pixel.
After a solution is obtained, the resulting attenuation variable, &agr;, for each pixel is mapped graphically, for example, by using darker colors for larger coefficients and lighter colors for smaller coefficients. The map obtained from this process is commonly called a reconstruction phantom. Since the attenuation coefficient of each pixel can be correlated with other physical parameters associated with the pixel, such as density (mass per unit volume), the mapping indicates the density pattern (or other parameter) within the phantom.
This type of tomography is called “continuous domain,” because values of the attenuation coefficients in the pixels are continuous, and pixelization can be made infinitesimally fine. Discrete domain tomography relates to ascertaining the presence or absence of objects by a single binary variable when the objects may be smaller than the beam irradiating them. Briefly, and according to the '175 patent, line counts are obtained from the crystal lattice under radiation representing the number of atoms in a specific row or line. A system of equations is derived from the line counts as described by the '175 patent and a solution obtained. Referring to
FIG. 13A
, the count of atoms in column C
1
(which is a line) is five. The line count of five produces the equation:
X
11
+X
21
+X
31
+X
41
+X
51
+X
61
=line count=5
In the equations, a value of “1” for an “X” indicates the presence of an atom and a value of “0” represents the absence. Then, a system of multiple equations is obtained from the multiple line counts that are solved by linear programming processes according to the '175 patent.
Detailed procedures for obtaining line counts are known in the art. One procedure is described in “Mapping Projected Potential, Interfacial Roughness, and Composition in General Crystalline Solids by Quantitative Electron Microscopy,” by P. Schwander et al.,
Phys. Rev. Lett.,
71, pp 4150-4153 (1993).
Referring to
FIG. 8
, a small sample
30
of a crystal is positioned in an electron beam
33
output by a transmission electron microscope, TEM. A detector detects the beam, and one can make inferences of physical features of the sample
30
. The Sample 30 of the '175 patent measured approximately 100 by 100 atoms in cross section. Larger samples in cross section, for example, 1000 atoms by 1000 atoms, and three dimensional samples, for example, slices of crystals of a size 25×1000×1000 atoms can also be analyzed. An expander
36
expands beam
33
. A detector
39
is, for example, a photographic film. The distance D is measurable by the human eye. When the beam
33
passes through an atom
5
, a relatively faint spot is produced on film
39
. When the beam passes through four atoms, the spot density may be intermediate in intensity, and so on. A ray
6
may encounter no atoms and produces no darkening. The gray levels of all spots are quantified and plotted producing a plot
42
(FIG.
12
). The electron beam
33
is moved horizontally across the sample at radius R per FIG.
9
.
The plot
42
can be made linear (although discrete) and of the form: y=mx+b, where y is the number of atoms, x is the grey scale value, m is the slope of a line (Delta) and b is the y-intercept. Correlations between grey scale level and number of atoms in a column are used in performing tomography on sample
30
. Complex patterns can be found as in FIG.
11
and linear plots formed as in
FIG. 12
or the sample may be rotated as in
FIG. 10
all to the result that linear programming processes may be mathematically applied to the detected line count data to obtain a reconstruction of the illuminated object
30
. For
Lee David
Vardi Yehuda
Bruce David V.
Lucent Technologies - Inc.
LandOfFree
Method and apparatus for discrete tomography does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for discrete tomography, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for discrete tomography will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2950513