US2003095696A1PendingUtilityA1
System, method and apparatus for small pulmonary nodule computer aided diagnosis from computed tomography scans
Priority: Sep 14, 2001Filed: Sep 16, 2002Published: May 22, 2003
Est. expirySep 14, 2021(expired)· nominal 20-yr term from priority
G06T 7/0012G06T 5/20G06T 7/11G06T 2207/10081A61B 6/5258G06T 2207/30064G06T 7/155
35
PatentIndex Score
0
Cited by
0
References
0
Claims
Abstract
The present invention is directed to diagnostic imaging of small pulmonary nodules. There are two main stages in the evaluation of pulmonary nodules from Computed Tomography (CT) scans: detection, in which the locations of possible nodules are identified, and characterization, in which a nodule is represented by measured features that may be used to evaluate the probability that the nodule is cancer. Currently, the most useful prediction feature is growth rate, which requires the comparison of size estimates from two CT scans recorded at different times. The present invention includes methods for detection and feature extraction for size characterization. The invention focuses the analysis of small pulmonary nodules that are less than 1 centimeter in size, but is also suitable for larger nodules as well.
Claims
exact text as granted — not AI-modifiedWhat is claimed is:
1 . A method for generating a lung mask for segmenting a lung image from voxel data containing the lung image, noise, solid components and surrounding background information, the method comprising:
(a) applying a median filter and a mean filter to the voxel data to reduce noise; (b) thresholding the noise reduced voxel data to identify solid components and other structures; (c) identifying the surrounding background in the noise reduced voxel data; (d) deleting the surrounding background from the noise reduced voxel data; (e) labeling connected components of the voxel data; (f) determining the largest connected components to select a lung region from the voxel data, the lung region having a geometric form; and (g) morphological filtering the voxel data to refine the geometric form of the lung region.
2 . The method as defined by claim 1 , wherein the voxel data is associated with a plurality of slices through a patient's lung, the slices beginning at about the patient's shoulders; and
step (a) is performed for the first 25 percent of the plurality of slices only whereby the computation time is substantially reduced.
3 . The method as defined by claim 1 , wherein the median filter has a size of 4×4.
4 . The method as defined by claim 1 , wherein the mean filter has a size of 1×3.
5 . The method as defined by claim 1 , wherein the voxel data in step (b) is thresholded at a gray level of about 500.
6 . The method as defined by claim 1 , wherein the largest connected components are associated with more than about 1 percent of the voxel data.
7 . The method as defined by claim 1 , wherein the voxel data is associated with a plurality of slices, the slices being divided into a first end region, a middle region, and a second end region; and
the morphological filtering being performed with 2D circular filter having a first diameter in the first end region and the second end region; and the morphological filtering being performed with 2D circular filter having a second diameter which is about twice the first diameter in the middle region.
8 . A method for measuring lung volume from a segmented lung image obtained from a scan which includes a plurality of slices of voxel data, the voxel data having a gray level value and a volume associated therewith, the method comprising:
generating a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice; determining a number of voxels in the segmented lung image from the matrix entries; and multiplying the number of voxels in the segmented lung image by the volume of each voxel.
9 . A method for measuring volume of tissue in a segmented lung image obtained from a scan including a plurality of slices of voxel data, the voxel data having a gray level value and a volume, the method comprising:
generating a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice; determining a sum of tissue of voxels in the segmented lung image from the matrix entries; and multiplying the sum of tissue of voxels in the segmented lung image by the volume of each voxel.
10 . The method as defined by claim 9 , wherein the sum of tissue of voxels is calculated by summing the product of each matrix entry multiplied by the corresponding gray level value divided by a gray level value assigned for tissue.
11 . A lung mask generating apparatus for segmenting a lung image from voxel data containing the lung image, noise, solid components and surrounding background information, the lung mask generating apparatus comprising:
a masking unit configured to:
(a) apply a median filter and a mean filter to the voxel data to reduce noise;
(b) threshold the noise reduced voxel data to identify solid components and other structures;
(c) identify the surrounding background in the noise reduced voxel data;
(d) delete the surrounding background from the noise reduced voxel data;
(e) label connected components of the voxel data;
(f) determine the largest connected components to select a lung region from the voxel data, the lung region having a geometric form; and
(g) morphologically filter the voxel data to refine the geometric form of the lung region.
12 . A lung mask generating apparatus as defined by claim 11 , wherein the voxel data is associated with a plurality of slices through a patient's lung, the slices beginning at about the patient's shoulders; and
step (a) is performed for the first 25 percent of the plurality of slices only whereby the computation time is substantially reduced.
13 . A lung mask generating apparatus as defined by claim 11 , wherein the median filter has a size of 4×4.
14 . A lung mask generating apparatus as defined by claim 11 , wherein the mean filter has a size of 1×3.
15 . A lung mask generating apparatus as defined by claim 11 , wherein the voxel data in step (b) is thresholded at a gray level of about 500.
16 . A lung mask generating apparatus as defined by claim 11 , wherein the largest connected components are associated with more than about 1 percent of the voxel data.
17 . A lung mask generating apparatus as defined by claim 11 , wherein the voxel data is associated with a plurality of slices, the slices being divided into a first end region, a middle region, and a second end region; and
the morphological filtering being performed with 2D circular filter having a first diameter in the first end region and the second end region; and the morphological filtering being performed with 2D circular filter having a second diameter which is about twice the first diameter in the middle region.
18 . An apparatus for measuring lung volume from a segmented lung image obtained from a scan which includes a plurality of slices of voxel data, the voxel data having a gray level value and a volume associated therewith, the apparatus comprising:
a lung volume measuring unit configured to:
generate a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice;
determine a number of voxels in the segmented lung image from the matrix entries; and
multiply the number of voxels in the segmented lung image by the volume of each voxel.
19 . An apparatus for measuring volume of tissue in a segmented lung image obtained from a scan including a plurality of slices of voxel data, the voxel data having a gray level value and a volume, the apparatus comprising:
a lung tissue measuring unit configured to:
generate a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice;
determine a sum of tissue of voxels in the segmented lung image from the matrix entries; and
multiply the sum of tissue of voxels in the segmented lung image by the volume of each voxel.
20 . An apparatus for measuring volume of tissue as defined by claim 19 , wherein the sum of tissue of voxels is calculated by summing the product of each matrix entry multiplied by the corresponding gray level value divided by a gray level value assigned for tissue.
21 . An article of manufacture for generating a lung mask for segmenting a lung image from voxel data containing the lung image, noise, solid components and surrounding background information, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
(a) applying a median filter and a mean filter to the voxel data to reduce noise;
(b) thresholding the noise reduced voxel data to identify solid components and other structures;
(c) identifying the surrounding background in the noise reduced voxel data;
(d) deleting the surrounding background from the noise reduced voxel data;
(e) labeling connected components of the voxel data;
(f) determining the largest connected components to select a lung region from the voxel data, the lung region having a geometric form; and
(g) morphological filtering the voxel data to refine the geometric form of the lung region.
22 . An article of manufacture for generating a lung mask as defined by claim 21 , wherein the voxel data is associated with a plurality of slices through a patient's lung, the slices beginning at about the patient's shoulders; and
step (a) is performed for the first 25 percent of the plurality of slices only whereby the computation time is substantially reduced.
23 . An article of manufacture for generating a lung mask as defined by claim 21 , wherein the median filter has a size of 4×4.
24 . An article of manufacture for generating a lung mask as defined by claim 21 , wherein the mean filter has a size of 1×3.
25 . An article of manufacture for generating a lung mask as defined by claim 21 , wherein the voxel data in step (b) is thresholded at a gray level of about 500.
26 . An article of manufacture for generating a lung mask defined by claim 21 , wherein the largest connected components are associated with more than about 1 percent of the voxel data.
27 . An article of manufacture for generating a lung mask as defined by claim 21 , wherein the voxel data is associated with a plurality of slices, the slices being divided into a first end region, a middle region, and a second end region; and
the morphological filtering being performed with 2D circular filter having a first diameter in the first end region and the second end region; and the morphological filtering being performed with 2D circular filter having a second diameter which is about twice the first diameter in the middle region.
28 . An article of manufacture for measuring lung volume from a segmented lung image obtained from a scan which includes a plurality of slices of voxel data, the voxel data having a gray level value and a volume associated therewith, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
generating a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice;
determining a number of voxels in the segmented lung image from the matrix entries; and
multiplying the number of voxels in the segmented lung image by the volume of each voxel.
29 . An article of manufacture for measuring volume of tissue in a segmented lung image obtained from a scan including a plurality of slices of voxel data, the voxel data having a gray level value and a volume, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
generating a matrix of entries from the segmented lung image, the matrix having a plurality of columns and a plurality of rows, each of the plurality of columns representing the gray level value and each of the plurality of rows representing one of the plurality of slices in the scan, the entries corresponding to the number of times that the gray level occurs in the corresponding slice;
determining a sum of tissue of voxels in the segmented lung image from the matrix entries; and
multiplying the sum of tissue of voxels in the segmented lung image by the volume of each voxel.
30 . An article of manufacture for measuring volume of tissue as defined by claim 29 , wherein the sum of tissue of voxels is calculated by summing the product of each matrix entry multiplied by the corresponding gray level value divided by a gray level value assigned for tissue.
31 . A method for finding the location, P′, and size, r′, of a pulmonary nodule in a high-resolution computed tomography (CT) image, the method comprising:
(a) selecting a set of initial processing parameters including an initial location, P 1 , an initial size, r 1 , and target value, T;
(b) computing an initial new location P i of the nodule with a locator template function;
(c) incrementally increasing size, r i ;
(d) computing a new location P i of the nodule with the locator template function;
(e) computing a size metric with a sizing template function;
(f) repeating steps (c) through (e) until the size metric is less than the target value;
(g) returning the location, P′, and the size, r′, from the previous iteration of steps (c) through (e).
32 . A method for finding the location, P, and size, r, of a pulmonary nodule in a high-resolution computed tomography image, the method comprising:
(a) windowing the image to ignore bone structures; (b) selecting a locator template function and a sizing template function; (c) selecting a set of initial processing parameters including: an initial location, P; size, r; and termination criteria; (d) performing a search to determine a maximum response of the locator template function; (e) determining a response of the sizing template function and comparing the response to the termination criteria; (f) incrementally increasing the size, r, only if the termination criteria has not been satisfied and repeating steps d and e; and (g) outputting the location, P, and size, r, of the nodule.
33 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the windowing comprises clipping the image at intensities over about 1000.
34 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the locator template function is a Gaussian template function.
35 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the locator template function is a Laplacian of the Gaussian template function.
36 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the locator template function is a difference of Gaussians template function.
37 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the template function has at least four parameters corresponding to the x-location, y-location, z-location, and radius.
38 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the initial location, P, is calculated from the image.
39 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the initial location, P, is specified by a user.
40 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the search is performed by a hill climbing method.
41 . A method for finding the location and size of a pulmonary nodule as defined in claim 32 , wherein the search is performed by Powell's method.
42 . A pulmonary nodule finding apparatus for finding the location, P′, and size, r′, of a pulmonary nodule in a high-resolution computed tomography (CT) image, the pulmonary nodule finding apparatus comprising:
a nodule finding unit configured to:
(a) select a set of initial processing parameters including an initial location, P 1 , an initial size, r 1 , and target value, T;
(b) compute an initial new location P i of the nodule with a locator template function;
(c) incrementally increase size, r i ;
(d) compute a new location P i of the nodule with the locator template function;
(e) compute a size metric with a sizing template function;
(f) repeat steps (c) through (e) until the size metric is less than the target value;
(g) return the location, P′, and the size, r′, from the previous iteration of steps (c) through (e).
43 . A pulmonary nodule finding apparatus for finding the location, P, and size, r, of a pulmonary nodule in a high-resolution computed tomography image, the pulmonary nodule finding apparatus comprising:
a nodule finding unit configured to:
(a) window the image to ignore bone structures;
(b) select a locator template function and a sizing template function;
(c) select a set of initial processing parameters including: an initial location, P; size, r; and termination criteria;
(d) perform a search to determine a maximum response of the locator template function;
(e) determine a response of the sizing template function and comparing the response to the termination criteria;
(f) incrementally increase the size, r, only if the termination criteria has not been satisfied and repeating steps d and e; and
(g) output the location, P, and size, r, of the nodule.
44 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the image is clipped at intensities over about 1000 to window the image.
45 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the locator template function is a Gaussian template function.
46 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the locator template function is a Laplacian of the Gaussian template function.
47 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the locator template function is a difference of Gaussians template function.
48 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the template function has at least four parameters corresponding to the x-location, y-location, z-location, and radius.
49 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the initial location, P, is calculated from the image.
50 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the initial location, P, is specified by a user.
51 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the search is performed by a hill climbing method.
52 . A pulmonary nodule finding apparatus as defined in claim 43 , wherein the search is performed by Powell's method.
53 . An article of manufacture for finding the location, P′, and size, r′, of a pulmonary nodule in a high-resolution computed tomography (CT) image, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
(a) selecting a set of initial processing parameters including an initial location, P 1 , an initial size, r 1 , and target value, T;
(b) computing an initial new location P i of the nodule with a locator template function;
(c) incrementally increasing size, r i ;
(d) computing a new location P i of the nodule with the locator template function;
(e) computing a size metric with a sizing template function;
(f) repeating steps (c) through (e) until the size metric is less than the target value;
(g) returning the location, P′, and the size, r′, from the previous iteration of steps (c) through (e).
54 . An article of manufacture for finding the location, P, and size, r, of a pulmonary nodule in a high-resolution computed tomography image, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
(a) windowing the image to ignore bone structures;
(b) selecting a locator template function and a sizing template function;
(c) selecting a set of initial processing parameters including: an initial location, P; size, r; and termination criteria;
(d) performing a search to determine a maximum response of the locator template function;
(e) determining a response of the sizing template function and comparing the response to the termination criteria;
(f) incrementally increasing the size, r, only if the termination criteria has not been satisfied and repeating steps d and e; and
(g) outputting the location, P, and size, r, of the nodule.
55 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the windowing comprises clipping the image at intensities over about 1000.
56 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the locator template function is a Gaussian template function.
57 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the locator template function is a Laplacian of the Gaussian template function.
58 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the locator template function is a difference of Gaussians template function.
59 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the template function has at least four parameters corresponding to the x-location, y-location, z-location, and radius.
60 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the initial location, P, is calculated from the image.
61 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the initial location, P, is specified by a user.
62 . An article of manufacture for finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the search is performed by a hill climbing method.
63 . An article of manufacture finding the location and size of a pulmonary nodule as defined in claim 54 , wherein the search is performed by Powell's method.
64 . A method for registering 3-d images of a pulmonary nodule from a high-resolution computed tomography (CT) scans, the images being in a floating point pixel-format associated with a 6-dimensional parameter space and including a first image (im 1 ) obtained at time-1 and a second image (im 2 ) obtained at time-2, the method comprising the steps of:
(a) calculating initial rigid-body transformation parameters for a rigid-body transformation on the first image (im 1 ); (b) determining the optimum rigid-body transformation parameters by calculating a registration metric between the second image (im 2 ) and the rigid-body transformation on the first image (im 1 ); and (c) generating a registered image from the optimum rigid-body transformation parameters.
65 . A method as defined in claim 64 , wherein step (a) is preceded by masking one of the images by setting pixels to a background value.
66 . A method as defined in claim 65 , wherein the background value is about −1000.
67 . A method as defined in claim 64 , wherein the registration metric is minimized.
68 . A method as defined in claim 64 , wherein the registration metric is maximized.
69 . A method as defined in claim 64 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t ); calculating the registration metric as a correlation (C) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the maximum correlation (C) in the 6-dimensional parameter space.
70 . A method as defined in claim 64 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t) ; calculating the registration metric as a mean-squared-difference (MSD) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the minimum mean-squared-difference (MSD) in the 6-dimensional parameter space.
71 . A method as defined in claim 69 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
ν
′
=
R
x
R
y
R
Z
ν
+
[
t
x
t
y
t
z
]
wherein R x , R y , and R z are rotation matrices defined as:
R
x
=
[
1
0
0
0
cos
(
r
x
)
-
sin
(
r
x
)
0
sin
(
r
x
)
cos
(
r
x
)
]
R
y
=
[
cos
(
r
y
)
0
sin
(
r
y
)
0
1
0
-
sin
(
r
y
)
0
cos
(
r
y
)
]
R
z
=
[
cos
(
r
z
)
-
sin
(
r
z
)
0
sin
(
r
z
)
cos
(
r
z
)
0
0
0
1
]
•
72 . A method as defined in claim 70 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
wherein R x , R y , and R z are rotation matrices defined as:
R x = [ 1 0 0 0 cos ( r x ) - sin ( r x ) 0 sin ( r x ) cos ( r x ) ] R y = [ cos ( r y ) 0 sin ( r y ) 0 1 0 - sin ( r y ) 0 cos ( r y ) ] R z = [ cos ( r z ) - sin ( r z ) 0 sin ( r z ) cos ( r z ) 0 0 0 1 ] •
73 . A method as defined in claim 64 , wherein initial rigid-body transformation parameters include six parameters (tx,ty,tz,rx,ry,rz) respectively defined as translation in x, translation in y, translation in z, rotation about the x-axis, rotation about the y-axis, and rotation about the z-axis;
wherein the initial rotation parameters (rx,ry,rz) are all set to zero; and the initial translation parameters (tx,ty,tz,) are set so that the nodule in the first image (im 1 ) overlaps the nodule in the second image (im 2 ) during the initial calculation of the registration metric.
74 . A method as defined in claim 73 , wherein the initial translation parameters (tx,ty,tz,) are set to a difference between the center of the first image (im 1 ) and the center of the second image (im 2 ).
75 . A method as defined in claim 73 , wherein the initial translation parameters (tx,ty,tz) are set to a difference between the center of mass of the first image (im 1 ) and the center of mass of the second image (im 2 ).
76 . A method as defined in claim 69 , wherein the searching is conducted by calculating the correlation (C) for every possible set of rigid-body transformation parameters.
77 . A method as defined in claim 70 , wherein the searching is conducted by calculating the mean-squared-difference (MSD) for every possible set of rigid-body transformation parameters.
78 . A method as defined in claim 69 , wherein the searching is conducted by a Hill-Climbing search method.
79 . A method as defined in claim 70 , wherein the searching is conducted by a Hill-Climbing search method.
80 . A method as defined in claim 69 , wherein the searching is conducted by Powell's method.
81 . A method as defined in claim 70 , wherein the searching is conducted by Powell's method.
82 . A registering apparatus for registering 3-d images of a pulmonary nodule from a high-resolution computed tomography (CT) scans, the images being in a floating point pixel-format associated with a 6-dimensional parameter space and including a first image (im 1 ) obtained at time-1 and a second image (im 2 ) obtained at time-2, the registering apparatus comprising:
a registering unit configured to:
(a) calculate initial rigid-body transformation parameters for a rigid-body transformation on the first image (im 1 );
(b) determine the optimum rigid-body transformation parameters by calculating a registration metric between the second image (im 2 ) and the rigid-body transformation on the first image (im 1 ); and
(c) generate a registered image from the optimum rigid-body transformation parameters.
83 . A registering apparatus as defined in claim 82 , wherein one of the images is initially masked by setting pixels to a background value.
84 . A registering apparatus as defined in claim 83 , wherein the background value is about −1000.
85 . A registering apparatus as defined in claim 82 , wherein the registration metric is minimized.
86 . A registering apparatus as defined in claim 82 , wherein the registration metric is maximized.
87 . A registering apparatus as defined in claim 82 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t ); calculating the registration metric as a correlation (C) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the maximum correlation (C) in the 6-dimensional parameter space.
88 . A registering apparatus as defined in claim 82 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t) ; calculating the registration metric as a mean-squared-difference (MSD) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the minimum mean-squared-difference (MSD) in the 6-dimensional parameter space.
89 . A registering apparatus as defined in claim 87 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
ν
′
=
R
x
R
y
R
Z
ν
+
[
t
x
t
y
t
z
]
wherein R x , R y , and R z are rotation matrices defined as:
R
x
=
[
1
0
0
0
cos
(
r
x
)
-
sin
(
r
x
)
0
sin
(
r
x
)
cos
(
r
x
)
]
R
y
=
[
cos
(
r
y
)
0
sin
(
r
y
)
0
1
0
-
sin
(
r
y
)
0
cos
(
r
y
)
]
R
z
=
[
cos
(
r
z
)
-
sin
(
r
z
)
0
sin
(
r
z
)
cos
(
r
z
)
0
0
0
1
]
•
90 . A registering apparatus as defined in claim 88 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
v
′
=
R
x
R
y
R
z
v
+
[
t
x
t
y
t
z
]
wherein R x , R y , and R z are rotation matrices defined as:
R
x
=
[
1
0
0
0
cos
(
r
x
)
-
sin
(
r
x
)
0
sin
(
r
x
)
cos
(
r
x
)
]
R
y
=
[
cos
(
r
y
)
0
sin
(
r
y
)
0
1
0
-
sin
(
r
y
)
0
cos
(
r
y
)
]
R
z
=
[
cos
(
r
z
)
-
sin
(
r
z
)
0
sin
(
r
z
)
cos
(
r
z
)
0
0
0
1
]
.
91 . A registering apparatus as defined in claim 82 , wherein initial rigid-body transformation parameters include six parameters (tx,ty,tz,rx,ry,rz) respectively defined as translation in x, translation in y, translation in z, rotation about the x-axis, rotation about the y-axis, and rotation about the z-axis;
wherein the initial rotation parameters (rx,ry,rz) are all set to zero; and the initial translation parameters (tx,ty,tz,) are set so that the nodule in the first image (im 1 ) overlaps the nodule in the second image (im 2 ) during the initial calculation of the registration metric.
92 . A registering apparatus as defined in claim 91 , wherein the initial translation parameters (tx,ty,tz,) are set to a difference between the center of the first image (im 1 ) and the center of the second image (im 2 ).
93 . A registering apparatus as defined in claim 91 , wherein the initial translation parameters (tx,ty,tz) are set to a difference between the center of mass of the first image (im 1 ) and the center of mass of the second image (im 2 ).
94 . A registering apparatus as defined in claim 87 , wherein the searching is conducted by calculating the correlation (C) for every possible set of rigid-body transformation parameters.
95 . A registering apparatus as defined in claim 88 , wherein the searching is conducted by calculating the mean-squared-difference (MSD) for every possible set of rigid-body transformation parameters.
96 . A registering apparatus as defined in claim 87 , wherein the searching is conducted by a Hill-Climbing search method.
97 . A registering apparatus as defined in claim 88 , wherein the searching is conducted by a Hill-Climbing search method.
98 . A registering apparatus as defined in claim 87 , wherein the searching is conducted by Powell's method.
99 . A registering apparatus as defined in claim 88 , wherein the searching is conducted by Powell's method.
100 . An article of manufacture for registering 3-d images of a pulmonary nodule from a high-resolution computed tomography (CT) scans, the images being in a floating point pixel-format associated with a 6-dimensional parameter space and including a first image (im 1 ) obtained at time-1 and a second image (im 2 ) obtained at time-2, the article comprising:
a machine readable medium containing one or more programs which when executed implement the steps of:
(a) calculating initial rigid-body transformation parameters for a rigid-body transformation on the first image (im 1 );
(b) determining the optimum rigid-body transformation parameters by calculating a registration metric between the second image (im 2 ) and the rigid-body transformation on the first image (im 1 ); and
(c) generating a registered image from the optimum rigid-body transformation parameters.
101 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein step (a) is preceded by masking one of the images by setting pixels to a background value.
102 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 101 , wherein the background value is about −1000.
103 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein the registration metric is minimized.
104 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein the registration metric is maximized.
105 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t ); calculating the registration metric as a correlation (C) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the maximum correlation (C) in the 6-dimensional parameter space.
106 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein the registration metric is calculated by
transforming the first image (im 1 ) with the initial rigid-body transformation parameters to obtain a transformed first image (im 1t) ; calculating the registration metric as a mean-squared-difference (MSD) between the transformed first image (im 1t ) and the second image (im 2 ); and searching for the minimum mean-squared-difference (MSD) in the 6-dimensional parameter space.
107 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 105 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
v
′
=
R
x
R
y
R
z
v
+
[
t
x
t
y
t
z
]
wherein R x , R y and R z are rotation matrices defined as:
R
x
=
[
1
0
0
0
cos
(
r
x
)
-
sin
(
r
x
)
0
sin
(
r
x
)
cos
(
r
x
)
]
R
y
=
[
cos
(
r
y
)
0
sin
(
r
y
)
0
1
0
-
sin
(
r
y
)
0
cos
(
r
y
)
]
R
z
=
[
cos
(
r
z
)
-
sin
(
r
z
)
0
sin
(
r
z
)
cos
(
r
z
)
0
0
0
1
]
.
108 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 106 , wherein the transforming of the first image (im 1 ) to obtain the transformed first image (im 1t ) is a mapping of a point v in 3-d space to a point v′ in transformed space defined by:
v
′
=
R
x
R
y
R
z
v
+
[
t
x
t
y
t
z
]
wherein R x , R y and R z are rotation matrices defined as:
R
x
=
[
1
0
0
0
cos
(
r
x
)
-
sin
(
r
x
)
0
sin
(
r
x
)
cos
(
r
x
)
]
R
y
=
[
cos
(
r
y
)
0
sin
(
r
y
)
0
1
0
-
sin
(
r
y
)
0
cos
(
r
y
)
]
R
z
=
[
cos
(
r
z
)
-
sin
(
r
z
)
0
sin
(
r
z
)
cos
(
r
z
)
0
0
0
1
]
.
109 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 100 , wherein initial rigid-body transformation parameters include six parameters (tx,ty,tz,rx,ry,rz) respectively defined as translation in x, translation in y, translation in z, rotation about the x-axis, rotation about the y-axis, and rotation about the z-axis;
wherein the initial rotation parameters (rx,ry,rz) are all set to zero; and the initial translation parameters (tx,ty,tz,) are set so that the nodule in the first image (im 1 ) overlaps the nodule in the second image (im 2 ) during the initial calculation of the registration metric.
110 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 109 , wherein the initial translation parameters (tx,ty,tz,) are set to a difference between the center of the first image (im 1 ) and the center of the second image (im 2 ).
111 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 109 , wherein the initial translation parameters (tx,ty,tz) are set to a difference between the center of mass of the first image (im 1 ) and the center of mass of the second image (im 2 ).
112 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 105 , wherein the searching is conducted by calculating the correlation (C) for every possible set of rigid-body transformation parameters.
113 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 106 , wherein the searching is conducted by calculating the mean-squared-difference (MSD) for every possible set of rigid-body transformation parameters.
114 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 105 , wherein the searching is conducted by a Hill-Climbing search method.
115 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 106 , wherein the searching is conducted by a Hill-Climbing search method.
116 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 105 , wherein the searching is conducted by Powell's method.
117 . An article of manufacture for registering 3-d images of a pulmonary nodule as defined in claim 106 , wherein the searching is conducted by Powell's method.
118 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule, the method comprising the steps of:
(a) providing an initial location P′; (b) calculating a spherical volume centered at the initial location P′, the spherical volume fitting inside the image; (c) calculating a center of mass COM of the spherical volume; (d) determining an initial direction d′, the initial direction d′ being directed towards the extraneous matter in accordance with the following equation d ′ = COM - P ′ COM - P ′ ; (e) initializing a current location P i to be equal to the initial location P′; (f) initializing a current direction di to be equal to the initial direction d′; (g) initializing a maximum ratio γ max , step size s, prior mass mass i-1 , and prior change in mass Δ i-1 ; (h) moving the current location P i by the step size s in the current direction d i ; (i) determining an equation defining a plane A, the plane A being normal to the current direction d i , the plane A passing through the current location P i ; (j) calculating a current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter; (k) calculating a current change in mass Δ i by subtracting the prior mass mass i-1 from the current mass mass i ; (l) calculating a current ratio γ in accordance with the following equation γ = Δ i Δ i - 1 - 1 ; (m) setting the prior mass mass i-1 equal to the current mass mass i ; (n) setting the prior change in mass Δ i-1 equal to the current change in mass Δ i ; (o) comparing the current ratio γ to the maximum ratio γ max ; (p) modifying the current direction d i to minimize the current mass mass i , and performing steps (h)-(o) while the current ratio γ is one of less than and equal to the maximum ratio γ max ; and (q) modifying the current direction d i , performing steps (h)-(o), and outputting the area of the nodule partitioned by the plane A in response to the current ratio γ being greater than the maximum ratio γ max .
119 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein the extraneous matter includes a pleural surface.
120 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein step (g) further includes the steps of:
initializing the maximum ratio γ max to 0.5; and initializing the step size s to 1.5.
121 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein step (j) further includes the steps of:
(r) defining the current location P i as being visited; (s) determining on which side of the plane A the current location P i is located; (t) terminating in reponse to the current location P i not being located on a side of the plane opposing that of the extraneous matter; (u) defining the current location P i as being part of a region of interest in response to the current location P i being located on the side of the plane opposing that of the extraneous matter; and (v) performing steps (r)-(u) recursively using a location corresponding to at least one of six (6) one-pixel moves from the current location P i .
122 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein step (p) further includes the steps of:
calculating recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s 1 from the current location P i ; and defining the current direction d i equal to the direction yielding the largest decrease in the current mass mass i .
123 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein step (q) further includes the steps of:
calculating recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s 1 from the current location P i ; and defining the current direction d i equal to the direction yielding the largest decrease in the current mass mass i .
124 . A method for removing extraneous matter from an image, the image including a juxtapleural nodule as defined in claim 118 , wherein the initial location P′ is located near a center of the nodule.
125 . A recursive apparatus for removing extraneous matter from an image, the image including a juxtapleural nodule, the recursive apparatus comprising:
a processing unit, the processing unit being configured to:
(a) accept an initial location P′;
(b) calculate a spherical volume centered at the initial location P′, the spherical volume fitting inside the image;
(c) calculate a center of mass COM of the spherical volume;
(d) determine an initial direction d′, the initial direction d′ being directed towards the extraneous matter in accordance with the following equation
d ′ = COM - P ′ COM - P ′ ;
(e) initialize a current location P i to be equal to the initial location P′;
(f) initialize a current direction d i to be equal to the initial direction d′;
(g) initialize a maximum ratio γ max , step size s, prior mass mass i-1 , and prior change in mass Δ i-1 ;
(h) move the current location P i by the step size s in the current direction d i ;
(i) determine an equation defining a plane A, the plane A being normal to the current direction d i , the plane A passing through the current location P i ;
(j) calculate a current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter;
(k) calculate a current change in mass Δ i by subtracting the prior mass mass i-1 from the current mass mass i ;
(l) calculate a current ratio γ in accordance with the following equation
γ = Δ i Δ i - 1 - 1 ;
(m) set the prior mass mass i-1 equal to the current mass mass i ;
(n) set the prior change in mass Δ i-1 equal to the current change in mass Δ i ;
(o) compare the current ratio γ to the maximum ratio γ max ;
(p) modify the current direction d i to minimize the current mass mass i , and performing steps (h)-(o) while the current ratio γ is one of less than and equal to the maximum ratio γ max ; and
(q) modify the current direction d i , performing steps (h)-(o), and outputting the area of the nodule partitioned by the plane A in response to the current ratio γ being greater than the maximum ratio γ max .
126 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein the extraneous matter includes a pleural surface.
127 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein in step (g) the processing unit is configured to:
initialize the maximum ratio γ max to 0.5; and initialize the step size s to 1.5.
128 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein in step (j) the processing unit is configured to:
(r) define the current location P i as being visited; (s) determine on which side of the plane A the current location P i is located; (t) terminate in reponse to the current location P i not being located on a side of the plane opposing that of the extraneous matter; (u) define the current location P i as being part of a region of interest in response to the current location P i being located on the side of the plane opposing that of the extraneous matter; and (v) perform steps (r)-(u) recursively using a location corresponding to at least one of six (6) one-pixel moves from the current location P i .
129 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein in step (p) the processing unit is configured to:
calculate recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s from the current location P t1 ; and define the current direction di equal to the direction yielding the largest decrease in the current mass mass i .
130 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein in step (q) the processing unit is configured to:
calculate recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s 1 from the current location P i ; and defining the current direction d i equal to the direction yielding the largest decrease in the current mass mass i .
131 . A recursive apparatus for removing extraneous matter from an image as defined in claim 125 , wherein the initial location P′ is located near a center of the nodule.
132 . An article of manufacture for removing extraneous matter from an image, the image including a juxtapleural nodule, the article comprising:
a machine readable medium containing at least one program which when executed implements the steps of:
(a) accepting an initial location P′;
(b) calculating a spherical volume centered at the initial location P′, the spherical volume fitting inside the image;
(c) calculating a center of mass COM of the spherical volume;
(d) determining an initial direction d′, the initial direction d′ being directed towards the extraneous matter in accordance with the following equation
d ′ = COM - P ′ || COM - P ′ || ;
(e) initializing a current location P i to be equal to the initial location P′;
(f) initializing a current direction d i to be equal to the initial direction d′;
(g) initializing a maximum ratio γ max , step size s, prior mass mass i-1 , and prior change in mass Δ i-1 ;
(h) moving the current location P i by the step size s in the current direction d i ;
(i) determining an equation defining a plane A, the plane A being normal to the current direction d i , the plane A passing through the current location P i ;
(j) calculating a current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter;
(k) calculating a current change in mass Δ i by subtracting the prior mass mass i-1 from the current mass mass i ;
(l) calculating a current ratio γ in accordance with the following equation
γ = Δ i Δ i - 1 - 1 ;
(m) setting the prior mass mass i-1 equal to the current mass mass i ;
(n) setting the prior change in mass Δ i-1 equal to the current change in mass Δ i ;
o) comparing the current ratio γ to the maximum ratio γ max ;
(p) modifying the current direction d i to minimize the current mass mass i , and performing steps (h)-(o) while the current ratio γ is one of less than and equal to the maximum ratio γ max ; and
(q) modifying the current direction d i , performing steps (h)-(o), and outputting the area of the nodule partitioned by the plane A in response to the current ratio γ being greater than the maximum ratio γ max .
133 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein the extraneous matter includes a pleural surface.
134 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein in step (g) the article further implements the steps of:
initializing the maximum ratio γ max to 0.5; and initializing the step size s to 1.5.
135 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein in step (j) the article further implements the steps of:
(r) defining the current location P i as being visited; (s) determining on which side of the plane A the current location P i is located; (t) terminating in reponse to the current location P i not being located on a side of the plane opposing that of the extraneous matter; (u) defining the current location P i as being part of a region of interest in response to the current location P i being located on the side of the plane opposing that of the extraneous matter; and (v) performing steps (r)-(u) recursively using a location corresponding to at least one of six (6) one-pixel moves from the current location P i .
136 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein in step (p) the article further implements the steps of:
calculating recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s 1 from the current location P i ; and defining the current direction d i equal to the direction yielding the largest decrease in the current mass mass i .
137 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein in step (q) the article further implements the steps of:
calculating recursively the current mass mass i of the nodule on a side of the plane A opposing that of the extraneous matter using at least one of six (6) directions and a step size s 1 from the current location P i ; and defining the current direction d i equal to the direction yielding the largest decrease in the current mass mass i .
138 . An article of manufacture for removing extraneous matter from an image as defined in claim 132 , wherein the initial location P′ is located near a center of the nodule.Cited by (0)
No later patents cite this yet.
References (0)
No backward citations on record.