Imaging system calibration
A system includes determination of a first sub-matrix of a projection matrix which describes a geometrical relationship between points of a three-dimensional coordinate system of the imaging system and points of a two-dimensional coordinate system of an image detector, determination of a second sub-matrix of the projection matrix, where the first and second sub-matrixes comprise a decomposition of the projection matrix, conversion of a first point of the two-dimensional coordinate system to a first point of the three-dimensional coordinate system based on the first and second sub-matrixes, determination of an updated first sub-matrix of an updated projection matrix, where the updated projection matrix describes a second geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and conversion of a second point of the two-dimensional coordinate system to a second point of the three-dimensional coordinate system based on the updated first and second sub-matrixes.
1. A calibration method for an imaging system comprising an imaging source and an image detector, comprising:
determining a first sub-matrix of a projection matrix, the projection matrix describing a geometrical relationship between points of a three-dimensional coordinate system of the imaging system and points of a two-dimensional coordinate system of the image detector;
determining a second sub-matrix of the projection matrix, where the first sub-matrix and the second sub-matrix comprise a decomposition of the projection matrix;
converting a first point of the two-dimensional coordinate system to a first point of the three-dimensional coordinate system based on the first sub-matrix and the second sub-matrix;
determining an updated first sub-matrix of an updated projection matrix according to a first calibration schedule, where the updated projection matrix describes a second geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the second sub-matrix comprise a decomposition of the updated projection matrix; and
converting a second point of the two-dimensional coordinate system to a second point of the three-dimensional coordinate system based on the updated first sub-matrix and the second sub-matrix.
2. The method according to claim 1 , further comprising:
determining an updated second sub-matrix of a second updated projection matrix according to a second calibration schedule, where the second calibration schedule is different from the first calibration schedule, where the second updated projection matrix describes a third geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the updated second sub-matrix comprise a decomposition of the second updated projection matrix; and
converting a third point of the two-dimensional coordinate system to a third point of the three-dimensional coordinate system based on the updated first sub-matrix and the updated second sub-matrix.
3. A method according to claim 1 ,
wherein determining the first sub-matrix comprises determining in-plane rotation φ of the image detector and determining a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source.
4. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a projection image of a predefined collimator leaf pattern;
determining one or more differences between the projection image of the predefined collimator leaf pattern and an expected projection image of the predefined collimator leaf pattern; and
determining in-plane rotation φ of the image detector and a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source based on the one or more differences.
5. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a first projection image of a predefined collimator leaf pattern with the collimator at a 0 degree position;
acquiring a second projection image of the predefined collimator leaf pattern with the collimator at a 90 degree position;
combining the first projection image and the second projection image to generate a superimposed image; and
determining in-plane rotation φ of the image detector and a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source based on the superimposed image.
6. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a projection image of a predefined collimator leaf pattern;
identifying an X r axis based on one or more locations of lateral edges of one or more leaves in the predefined collimator leaf pattern; and
identifying a Y r axis based on one or more locations of longitudinal edges of one or more leaves in the predefined collimator leaf pattern.
7. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a projection image of a predefined collimator leaf pattern in which edges of one or more leaves are symmetrically located about the X r and Y r axes;
identifying a first one or more lines defined by lateral edges of one or more leaves in the projection image;
identifying a first one or more lines defined by longitudinal edges of one or more leaves in the projection image;
identifying an axis of symmetry of the first one or more lines as the X r axis; and
identifying an axis of symmetry of the second one or more lines as the Y r axis.
8. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a projection image of a collimator comprising two opposing Y-jaws, in which each Y-jaw defines a notch at its lateral center;
identifying each notch in the projection image; and
identifying the Y r axis as a line joining the notches.
9. The method according to claim 3 , wherein determining the first sub-matrix comprises:
acquiring a projection image of a collimator in which a first pair of opposing leaves located at an upper Y boundary of the collimator define a first narrow gap, and a second pair of opposing leaves located at a lower Y boundary of the collimator define a second narrow gap;
identifying the first narrow gap and the second narrow gap in the projection image; and
identifying the Y r axis as a line joining the first narrow gap and the second narrow gap.
10. The method according to claim 1 , wherein the first sub-matrix is determined as
P
θ
Flexible
=
⌈
1
p
w
0
u
0
0
-
1
p
h
v
0
0
0
-
1
⌉
*
[
cos
∅
-
sin
∅
0
sin
∅
cos
∅
0
0
0
1
]
θ
︸
P
θ
Flexible
︷
P
θ
,
wherein the second sub-matrix is determined as
P
θ
Rigid
=
[
-
f
0
0
0
0
-
f
0
0
0
0
0
1
]
θ
*
[
T
]
θ
*
[
R
]
θ
and
wherein the projection matrix is P θ =P θ Flexible *P θ Rigid , where Pw and Ph represent pixel dimensions on the FPD, θ represents a gantry angle, [T] θ represents translation quantity, [R] θ represents rotation quantity, and f represents a distance between the imaging source and the image detector.
11. The method according to claim 1 , further comprising:
acquiring a plurality of projection images of a plurality of fiducials at respective different gantry angles;
generating a three-dimensional image of the plurality of fiducials based on the plurality of projection images;
determining respective dimensions of the plurality of fiducials based on the three-dimensional image;
determining whether the determined dimensions differ from actual sizes of the plurality of fiducials by more than a threshold; and
if the determined sizes differ from actual dimensions of the plurality of fiducials by more than the threshold:
determining an updated second sub-matrix of a second updated projection matrix, where the second updated projection matrix describes a third geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the updated second sub-matrix comprise a decomposition of the second updated projection matrix; and
converting a third point of the two-dimensional coordinate system to a third point of the three-dimensional coordinate system based on the updated first sub-matrix and the updated second sub-matrix.
12. The method according to claim 11 , wherein the plurality of fiducials comprise a plurality of spherical radio-opaque ball bearings attached to a treatment couch, and
wherein determining respective dimensions of the plurality of fiducials comprises determining blur of each of the plurality of fiducials based on the three-dimensional image.
13. An imaging system comprising: an imaging source to emit radiation;
an image detector to receive the radiation and to generate images based on the received radiation; and
a processing unit configured to execute program code stored on a non-transitory computer readable medium to cause the system to:
determine a first sub-matrix of a projection matrix, the projection matrix describing a geometrical relationship between points of a three-dimensional coordinate system of the imaging system and points of a two-dimensional coordinate system of the image detector;
determine a second sub-matrix of the projection matrix, where the first sub-matrix and the second sub-matrix comprise a decomposition of the projection matrix;
convert a first point of the two-dimensional coordinate system to a first point of the three-dimensional coordinate system based on the first sub-matrix and the second submatrix;
determine an updated first sub-matrix of an updated projection matrix according to a first calibration schedule, where the updated projection matrix describes a second geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the second sub-matrix comprise a decomposition of the updated projection matrix; and
convert a second point of the two-dimensional system to a second point of the three-dimensional coordinate system based on the updated first sub-matrix and the second sub-matrix.
14. The system according to claim 13 , the processing unit to further execute program code stored on a non-transitory computer readable medium to cause the system to:
determine an updated second sub-matrix of a second updated projection matrix according to a second calibration schedule, where the second calibration schedule is different from the first calibration schedule, where the second updated projection matrix describes a third geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the updated second sub-matrix comprise a decomposition of the second updated projection matrix; and
convert a third point of the two-dimensional coordinate system to a third point of the three-dimensional coordinate system based on the updated first sub-matrix and the updated second sub-matrix.
15. The system according to claim 13 ,
wherein determination of the first sub-matrix comprises determination of in-plane rotation φ of the image detector and determination of a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source.
16. The system according to claim 15 , further comprising a collimator, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to acquire a projection image of a predefined leaf pattern of the collimator;
determination of one or more difference between the projection image of the predefined leaf pattern and an expected projection image of the predefined leaf pattern; and
determination of in-plane rotation φ of the image detector and a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source based on the one or more differences.
17. The system according to claim 15 , further comprising a collimator, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to acquire a first projection image of a predefined leaf pattern of the collimator with the collimator at a 0 degree position;
operation of the imaging source and the imaging detector to acquire a second projection image of a predefined leaf pattern of the collimator with the collimator at a 90 degree position;
combination of the first projection image and the second projection image to generate a superimposed image; and
determination of in-plane rotation φ of the image detector and a point of the two-dimensional coordinate system which is intercepted by a beam axis of the imaging source based on the superimposed image.
18. The system according to claim 15 , further comprising a collimator, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to acquire a projection image of a predefined leaf pattern of the collimator;
identification of an X r axis based on one or more locations of lateral edges of one or more leaves in the predefined leaf pattern; and
identification of a Y r axis based on one or more locations of longitudinal edges of one more leaves in the predefined leaf pattern.
19. The system according to claim 15 , further comprising a collimator, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to acquire a projection image of a predefined leaf pattern of the collimator in which edges of one or more leaves are symmetrically located about the X r and Y r axes;
identification of a first one or more lines defined by lateral edges of one or more leaves in the projection image;
identification of a first one or more lines defined by longitudinal edges of one or more leaves in the projection image;
identification of an axis of symmetry of the first one or more lines as the X r axis; and
identification of an axis of symmetry of the second one or more lines as the Y r axis.
20. The system according to claim 15 , further comprising a collimator comprising two opposing Y-jaws, in which each Y-jaw defines a notch at its lateral center, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to acquire a projection image of the collimator;
identification of each notch in the projection image; and
identification of the Y r axis as a line joining the notches.
21. The system according to the claim 15 , further comprising a collimator in which a first pair of opposing leaves located at an upper Y boundary of the collimator define a first narrow gap, and a second pair of opposing leaves located at a lower Y boundary of the collimator define a second narrow gap, and wherein determination of the first sub-matrix comprises:
operation of the imaging source and the imaging detector to a projection image of a collimator;
identification of the first narrow gap and the second narrow gap in the projection image; and
identification of the Y r axis as a line joining the first narrow gap and the second narrow gap.
22. The system according to claim 13 , wherein the first sub-matrix is determined as
P
θ
Flexible
=
⌈
1
p
w
0
u
0
0
-
1
p
h
v
0
0
0
-
1
⌉
*
[
cos
∅
-
sin
∅
0
sin
∅
cos
∅
0
0
0
1
]
θ
︸
P
θ
Flexible
︷
P
θ
,
wherein the second sub-matrix is determined as
P
θ
Rigid
=
[
-
f
0
0
0
0
-
f
0
0
0
0
0
1
]
θ
*
[
T
]
θ
*
[
R
]
θ
and
wherein the projection matrix is P θ =P θ Flexible *P θ Rigid , where Pw and Ph represent pixel dimensions on the FPD, θ represents a gantry angle, [T] θ represents translation quantity, [R] θ represents rotation quantity, and f represents a distance between the imaging source and the image detector.
23. The system according to claim 13 , the processing unit to further execute
program code stored on a non-transitory computer readable medium to cause the system to:
operate the imaging source and the imaging detector to acquire a plurality of projection images of a plurality of fiducials at respective different gantry angles;
generate a three-dimensional image of the plurality of fiducials based on the plurality of projection images;
determine respective dimensions of the plurality of fiducials based on the three-dimensional images;
determine whether the determined dimensions differ from actual sizes of the plurality of fiducials by more than a threshold; and
if the determined sizes differ from actual dimensions of the plurality of fiducials by more than the threshold:
determining an updated second sub-matrix of a second updated projection matrix, where the second updated projection matrix describes a third geometrical relationship between points of the three-dimensional coordinate system and points of the two-dimensional coordinate system, and where the updated first sub-matrix and the updated second sub-matrix comprise a decomposition of the second updated projection matrix; and
converting a third point of the two-dimensional coordinate system to a third point of the three-dimensional coordinate system based on the updated first sub-matrix and the updated second sub-matrix.
24. The system according to claim 23 , wherein the plurality of fiducials comprise a plurality of spherical radio-opaque ball bearings attached to a treatment couch, and
wherein determining respective dimensions of the plurality of fiducials comprises determining of blur of each of the plurality of fiducials based on the three-dimensional image.