CA2452528A1 - System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in ct imaging - Google Patents
System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in ct imaging Download PDFInfo
- Publication number
- CA2452528A1 CA2452528A1 CA 2452528 CA2452528A CA2452528A1 CA 2452528 A1 CA2452528 A1 CA 2452528A1 CA 2452528 CA2452528 CA 2452528 CA 2452528 A CA2452528 A CA 2452528A CA 2452528 A1 CA2452528 A1 CA 2452528A1
- Authority
- CA
- Canada
- Prior art keywords
- illumination
- function
- counter
- imaging data
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000005286 illumination Methods 0.000 title claims abstract description 53
- 238000000034 method Methods 0.000 title claims description 30
- 238000003384 imaging method Methods 0.000 title claims description 24
- 230000003044 adaptive effect Effects 0.000 claims abstract description 22
- 238000002591 computed tomography Methods 0.000 claims abstract description 17
- 230000000916 dilatatory effect Effects 0.000 claims 2
- 239000002184 metal Substances 0.000 abstract description 8
- 238000001914 filtration Methods 0.000 abstract description 7
- 239000007943 implant Substances 0.000 abstract description 6
- 238000012937 correction Methods 0.000 abstract description 3
- 238000009877 rendering Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 4
- 210000001519 tissue Anatomy 0.000 description 4
- 230000010339 dilation Effects 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 210000000988 bone and bone Anatomy 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 241000284156 Clerodendrum quadriloculare Species 0.000 description 1
- 235000000177 Indigofera tinctoria Nutrition 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000000593 degrading effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 229940097275 indigo Drugs 0.000 description 1
- COHYTHOBJLSHDF-UHFFFAOYSA-N indigo powder Natural products N1C2=CC=CC=C2C(=O)C1=C1C(=O)C2=CC=CC=C2N1 COHYTHOBJLSHDF-UHFFFAOYSA-N 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 210000000689 upper leg Anatomy 0.000 description 1
Classifications
-
- G06T5/70—
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5258—Devices using data or image processing specially adapted for radiation diagnosis involving detection or reduction of artifacts or noise
-
- G06T5/92—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Abstract
In computed tomography images (102), streak artifacts caused by the presence of metal implants and inhomogeneous estimation of tissue density are reduced or eliminated. The algorithm has two basic steps: 1)illumination correction (104) and 2) adaptive three-Dimensional filtering (106). The algorithm start s by estimating the direction of the streak and the degree of inhomogeneous densities by gray scale morphology dilatation. Then, it proceeds to estimate the correct densities based on the estimations and to reduce the streak by a n adaptive three-dimensional filtering (106) whose parameters depend on the streak direction and the local image contrast.
Description
SYSTEM AND METHOD FOR DEDUCING OR ELIMINATING STREAK
ARTIFACTS AND ILLUMINATION INHOMOGENEITY IN CT IMAGING
Field of the Invention The present invention is directed to a system and method for processing images obtained through computed tomography (CT) or the like and more particularly to such a system and method in which artifacts caused by the presence of metal implants are reduced or eliminated.
Description of Related Art X-ray computed tomography (CT) is an important imaging technique, particularly in 1o medical and dental applications. A series of X-ray beams from many different angles are used to create cross-sectional images of the patient's body. Those images from multiple slices are assembled in a computer into a three-dimensional picture that can display organs, bones, and tissues in great detail. CT offers high spatial resolution, three-dimensional registration and minimal blurnng caused by motion.
However, the presence of strongly attenuating objects, such as metal implants or fillings, causes streak artifacts, also called starburst artifacts, in the image. Another problem encountered in CT is inhomogeneous estimation of tissue density caused by inhomogeneous illumination. While much investigation has been done into reducing or eliminating those problems, a satisfactory technique has not yet been found.
One attempt at a solution to the problem of artifacts caused by metal implants is found in U.S. Patent No. 5,561,695 to Hu. That patent teaches a method for improving CT image quality in which data from a helical reconstruction are separated into a background component and a sharp-structure component. The separation can be performed using gray-scale thresholding, since the sharp structures and the image background usually have widely differing CT numbers. The image background is filtered to remove high-frequency artifacts.
The images are recombined. However, the technique of Hu introduces more computational complexity than is desired, and the Hu approach does not appear to have been widely adopted by the radiology community.
Summary of the Invention It will be readily apparent from the above that a need exists in the art for a computationally efficient technique for reducing or eliminating streak artifacts and illumination inhomogeneity.
It is therefore an object of the present invention to reduce or eliminate streak artifacts caused by metal implants or the like.
It is another object of the present invention to correct illumination inhomogeneities.
It is still another object of the present invention to achieve the above objects in a computationally efficient way.
to To achieve that and other objects, the present invention is directed to a technique for CT or other imaging that works directly with the reconstructed data. The technique has two basic steps: 1) illumination correction and 2) adaptive 3D filtering. The algorithm starts by estimating the direction of the streak and the degree of inhomogeneous densities by gray scale morphology dilation. Then, it proceeds to estimate the correct densities based on the estimations and to reduce the streak by an adaptive 3D filtering whose parameters depend on the streak direction and the local image contrast.
Brief Description of the Drawings A preferred embodiment of the present invention will be set forth in detail with reference to the drawings, in which:
Figs. 1-3 show flow charts of the process according to the preferred embodiment;
Figs. 4A-4E show steps in the processing of a slice of an image;
Figs. 5A and SB show volume renderings of the image from raw data and processed data, respectively; and Fig. 6 shows a block diagram of a system on which the preferred embodiment can be implemented.
1o Detailed Description of the Preferred Embodiment The preferred embodiment of the present invention will now be set forth in detail with reference to the drawings.
Fig. 1 shows an overview of the process carried out in the preferred embodiment.
A$er the raw data have been taken in step 102, the process includes two steps:
the illumination correction of step 104 and the adaptive 3D filtering of step 106.
Those two steps will be explained in detail.
From a set of reconstructed CT images f(x, y) that contain a metal implant, the algorithm proceeds to stack those images in a three dimensional (3D) volumetric image f(x, y, z). Then, the algorithm gray scale dilates every single slice to estimate a propagation potential field p(x, y, z) of the streak artifacts and to estimate the inhomogeneous illumination I(x, y, z).
The estimation of the potential field p(x, y, z) via gray scale dilation is performed through the following steps shown in Fig. 2:
Step 202. Create a copy of the raw data: r(x, y, z) = f(x, y, z) Step 204. Initialize p(x, y, z) = 0; set counter c to 1.
Step 206. Gray scale dilate every reconstructed CT slice image f(x, y):
g(x, y) = r(x, y) D h(x, y) where h(x, y) is the kernel h(x, y) = 1 1 1 2o for odd values of c and h(x, y) = 1 1 1 for even values of c. The dilation process establishes a growing front from the regions of high intensity. .
Step 208. Set p(x, y, z) = c and estimate the derivative of the illumination function I(x, y, z) along the growing path for all the points where g(x, y, z) > r(x, y, z).
From the directional derivative, reconstruct the illumination function. That step will be described below with reference to Fig. 3.
Step 210. Set r(x, y, z) = g(x, y, z).
Step 212. Increment c: c = c + 1.
Step 214. Determine whether any points in p(x, y, z) have changes in value. If so, l0 steps 206-212 are repeated. Otherwise, the process ends in step 216.
The reconstruction of the illumination function, mentioned above with reference to step 208, will now be described. The metal inside the body will introduce some changes in the illumination profile of the image. It is assumed that the image is composed of piece-wise elements, each one with uniform density. Therefore, there should not be any strong variations on the image density inside any tissue, and the illumination function can be reconstructed by integrating the estimation of the illumination derivative, ~c 1(x, y, z), in the direction of the propagation front. The estimation of a I(x, y, z) is done by an adaptive weighted difference a~
of the neighborhood differences and the previous neighborhood estimations of a I(x, y, z) .
ac If ~ I(x, y, z) is initialized to zero in step 202 of the propagation front estimation, 2o then in step 208, the directional derivative is estimated using the following formulation:
_a 8c I (x~ Y~ z) -> >
a ~ ~ (~(i, j, c)[(1.0 - y(i, j))( f (x + i, y + j, z) - f (x, y, z)) + y(i, j)1 p (x + i, y + j, z)]) i°_1 l-_1 where 1 if g(x + i, y + j, z) > r(x + i, y + j, z) ~(Z ~ J s C) - a (ac) 0 otherwise a is a normalization constant such that ue~'~~z EE~(i, j, c) =1, and Y(Z ~ J) = 1 if ~ log[I .f (x + i, y + j, z) - f (x, y, z) ~ ~ 7c I (x + i, y + j, z)~ ~~ ~
0 otherwise The above equations provide an estimation of the illumination directional derivative at all points where p(x, y, z) = c.
The quantity 6 is the discontinuity threshold and depends on the body being imaged.
For human studies, a good threshold is a = 0.14. On the other hand, the constant a depends on the image resolution. For a voxel resolution of 0.5 x 0.5 mm2, a good value is a = 0.03. It will be seen from the above that the weighting function w(i, j, c) depends on the growing front, while y(i, j) depends on the image discontinuities.
Once the derivative of the illumination function has been estimated, it is integrated to get the illumination function. The integration starts by setting I(x, y, z) =
k, where k is a constant equal to the mean image intensity. Once I(x, y, z) has been initialized, the integration starts from the outer boundary of the image to the point where the potential image is zero. The process is carried out through the following steps shown in Fig.
3:
Step 302. Initialize I(x, y, z) = k set counter c to the maximum potential, c = max(p(x, Y. z)).
Step 304. Look at the neighborhood of all the points where p(x, y, z) = c; at those points, I(x,y,z)=u~~~,(3(i,j)[~ I(x+i,y+j,z)+I(x,y,z)J, -_y=
where 1 if p(x + i, y + j, z) > p(x, y, z) 0 otherwise and a is a normalization constant such that uEE(3(i, j) = 1.
Step 306. Decrement c: c = c - 1.
Step 308. Determine whether c < 1. Repeat steps 304 and 306 until c < 1, at which time the process ends in step 310.
Once the illumination field is found, the illumination inhomogeneity in f(x, y, z) is corrected:
g(x~ Y~ z) = f (x~ Y~ z)k I (x, y, z) The new corrected image g(x, y, z) has smaller variations in the image intensity than the original image, particularly at those points at which the density should be constant.
In the real world the original image has been degraded in such a way that the discontinuity detection can be very hard. To avoid that problem, the image f(x, y, z) is pre-filtered with a non-linear structure preserving filter that reduced the noise around some of the streak artifacts. The filtered image was used for the estimation of the illumination profile.
Once the image has been corrected as described above, the streak artifact can be reduced. A 3 x 3 x 3 adaptive filter L(x, y, z) is used to reduce the streak artifact. The filter coefficient is a function if the propagation potential and the local image contrast and is given by L(x,y,z) _ (s r~) (s+(0.1+(p(xo +x, Yo '+' Y~zo +z)-P(xo~Yo~zo))Z)(g(xo '~' x~Yo +Y~zo ~' z)-g(xo~Yo~zo))Z) 2o where s is the noise estimation variance at point (x0, yo, zo) and r1 is the normalization constant.
Coefficients at neighboring pomis with different potential as well as those neighboring points whose density is very different will be very small. On the other hand, coefficients at neighboring points with similar potential and very similar density compared to the noise estimation variance are larger. This formulation reduces artifacts and preserves points where is a strong contrast between tissues.
The final processed image h(xo, yo, zo) at the point (xo, yo, zo) is given by h(xo ~ Yo ~ zo ) - ~ ~ ~ L(xo + x~ Yo + Y~ zo + z)g(xo + x~ Yo + Y~ zo + z) .
X-_y-_IZ-_I
The adaptive filtering process can be done several times so that it effectively removes most of the streaks present in the image.
1o In cases where the in-plane resolution is much finer than the slice thickness, the adaptive filter can be modified in such a way as to avoid degrading the image.
The image quality can also be improved by providing a space and streak oriented noise estimation. If that way, those regions are filtered where the streak artifacts are more important, and filtering can be avoided in those regions where the streaks are not so strong.
The algorithm has been tested on several CT images with a hip prosthesis. Fig.
shows a single slice of one of those images. The image slice contains the femur stem and the prosthesis cup. Fig. 4B shows the estimated potential field, while Fig. 4C
shows the estimation of the illumination field. Fig. 4D is the CT image after removing the illumination artifacts. Fig. 4E shows the filtered image after smoothing the artifacts with the adaptive filter. Figs. 5A and SB show the volume rendering of the hip from the raw data and the processed data, respectively. As one can see the volumetric rendering of the hip from the processed data allows to see the metal prosthesis as well as more detail in the bone that surrounds the prosthesis.
The embodiment disclosed above and other embodiments can be implemented in a system such as that shown in the block diagram of Fig. 6. The system 600 includes an input device 602 for input of the image data and the like. The information input through the input device 602 is received in the workstation 604, which has a storage device 606 such as a hard drive, a processing unit 608 for performing the processing disclosed above, and a graphics rendering engine 610 for preparing the final processed image for viewing, e.g., by surface rendering. An output device 612 can include a monitor for viewing the images rendered by the rendering engine 610, a further storage device such as a video recorder for recording the images, or both. Illustrative examples of the workstation 604 and the graphics rendering engine 610 are a Silicon Graphics Indigo workstation and an Irix Explorer 3D
graphics engine.
While a preferred embodiment has been set forth above in detail, those skilled in the art who have reviewed the present disclosure will readily appreciate that other embodiments can be realized within the scope of the invention. For example, numerical values, such as those given for a and 6, are illustrative rather than limiting. Also, the imaged object can be any human or animal body part or any other object. Moreover, the invention has applicability to imaging technologies other than CT. Therefore, the present invention should be construed as limited only by the appended claims.
to
ARTIFACTS AND ILLUMINATION INHOMOGENEITY IN CT IMAGING
Field of the Invention The present invention is directed to a system and method for processing images obtained through computed tomography (CT) or the like and more particularly to such a system and method in which artifacts caused by the presence of metal implants are reduced or eliminated.
Description of Related Art X-ray computed tomography (CT) is an important imaging technique, particularly in 1o medical and dental applications. A series of X-ray beams from many different angles are used to create cross-sectional images of the patient's body. Those images from multiple slices are assembled in a computer into a three-dimensional picture that can display organs, bones, and tissues in great detail. CT offers high spatial resolution, three-dimensional registration and minimal blurnng caused by motion.
However, the presence of strongly attenuating objects, such as metal implants or fillings, causes streak artifacts, also called starburst artifacts, in the image. Another problem encountered in CT is inhomogeneous estimation of tissue density caused by inhomogeneous illumination. While much investigation has been done into reducing or eliminating those problems, a satisfactory technique has not yet been found.
One attempt at a solution to the problem of artifacts caused by metal implants is found in U.S. Patent No. 5,561,695 to Hu. That patent teaches a method for improving CT image quality in which data from a helical reconstruction are separated into a background component and a sharp-structure component. The separation can be performed using gray-scale thresholding, since the sharp structures and the image background usually have widely differing CT numbers. The image background is filtered to remove high-frequency artifacts.
The images are recombined. However, the technique of Hu introduces more computational complexity than is desired, and the Hu approach does not appear to have been widely adopted by the radiology community.
Summary of the Invention It will be readily apparent from the above that a need exists in the art for a computationally efficient technique for reducing or eliminating streak artifacts and illumination inhomogeneity.
It is therefore an object of the present invention to reduce or eliminate streak artifacts caused by metal implants or the like.
It is another object of the present invention to correct illumination inhomogeneities.
It is still another object of the present invention to achieve the above objects in a computationally efficient way.
to To achieve that and other objects, the present invention is directed to a technique for CT or other imaging that works directly with the reconstructed data. The technique has two basic steps: 1) illumination correction and 2) adaptive 3D filtering. The algorithm starts by estimating the direction of the streak and the degree of inhomogeneous densities by gray scale morphology dilation. Then, it proceeds to estimate the correct densities based on the estimations and to reduce the streak by an adaptive 3D filtering whose parameters depend on the streak direction and the local image contrast.
Brief Description of the Drawings A preferred embodiment of the present invention will be set forth in detail with reference to the drawings, in which:
Figs. 1-3 show flow charts of the process according to the preferred embodiment;
Figs. 4A-4E show steps in the processing of a slice of an image;
Figs. 5A and SB show volume renderings of the image from raw data and processed data, respectively; and Fig. 6 shows a block diagram of a system on which the preferred embodiment can be implemented.
1o Detailed Description of the Preferred Embodiment The preferred embodiment of the present invention will now be set forth in detail with reference to the drawings.
Fig. 1 shows an overview of the process carried out in the preferred embodiment.
A$er the raw data have been taken in step 102, the process includes two steps:
the illumination correction of step 104 and the adaptive 3D filtering of step 106.
Those two steps will be explained in detail.
From a set of reconstructed CT images f(x, y) that contain a metal implant, the algorithm proceeds to stack those images in a three dimensional (3D) volumetric image f(x, y, z). Then, the algorithm gray scale dilates every single slice to estimate a propagation potential field p(x, y, z) of the streak artifacts and to estimate the inhomogeneous illumination I(x, y, z).
The estimation of the potential field p(x, y, z) via gray scale dilation is performed through the following steps shown in Fig. 2:
Step 202. Create a copy of the raw data: r(x, y, z) = f(x, y, z) Step 204. Initialize p(x, y, z) = 0; set counter c to 1.
Step 206. Gray scale dilate every reconstructed CT slice image f(x, y):
g(x, y) = r(x, y) D h(x, y) where h(x, y) is the kernel h(x, y) = 1 1 1 2o for odd values of c and h(x, y) = 1 1 1 for even values of c. The dilation process establishes a growing front from the regions of high intensity. .
Step 208. Set p(x, y, z) = c and estimate the derivative of the illumination function I(x, y, z) along the growing path for all the points where g(x, y, z) > r(x, y, z).
From the directional derivative, reconstruct the illumination function. That step will be described below with reference to Fig. 3.
Step 210. Set r(x, y, z) = g(x, y, z).
Step 212. Increment c: c = c + 1.
Step 214. Determine whether any points in p(x, y, z) have changes in value. If so, l0 steps 206-212 are repeated. Otherwise, the process ends in step 216.
The reconstruction of the illumination function, mentioned above with reference to step 208, will now be described. The metal inside the body will introduce some changes in the illumination profile of the image. It is assumed that the image is composed of piece-wise elements, each one with uniform density. Therefore, there should not be any strong variations on the image density inside any tissue, and the illumination function can be reconstructed by integrating the estimation of the illumination derivative, ~c 1(x, y, z), in the direction of the propagation front. The estimation of a I(x, y, z) is done by an adaptive weighted difference a~
of the neighborhood differences and the previous neighborhood estimations of a I(x, y, z) .
ac If ~ I(x, y, z) is initialized to zero in step 202 of the propagation front estimation, 2o then in step 208, the directional derivative is estimated using the following formulation:
_a 8c I (x~ Y~ z) -> >
a ~ ~ (~(i, j, c)[(1.0 - y(i, j))( f (x + i, y + j, z) - f (x, y, z)) + y(i, j)1 p (x + i, y + j, z)]) i°_1 l-_1 where 1 if g(x + i, y + j, z) > r(x + i, y + j, z) ~(Z ~ J s C) - a (ac) 0 otherwise a is a normalization constant such that ue~'~~z EE~(i, j, c) =1, and Y(Z ~ J) = 1 if ~ log[I .f (x + i, y + j, z) - f (x, y, z) ~ ~ 7c I (x + i, y + j, z)~ ~~ ~
0 otherwise The above equations provide an estimation of the illumination directional derivative at all points where p(x, y, z) = c.
The quantity 6 is the discontinuity threshold and depends on the body being imaged.
For human studies, a good threshold is a = 0.14. On the other hand, the constant a depends on the image resolution. For a voxel resolution of 0.5 x 0.5 mm2, a good value is a = 0.03. It will be seen from the above that the weighting function w(i, j, c) depends on the growing front, while y(i, j) depends on the image discontinuities.
Once the derivative of the illumination function has been estimated, it is integrated to get the illumination function. The integration starts by setting I(x, y, z) =
k, where k is a constant equal to the mean image intensity. Once I(x, y, z) has been initialized, the integration starts from the outer boundary of the image to the point where the potential image is zero. The process is carried out through the following steps shown in Fig.
3:
Step 302. Initialize I(x, y, z) = k set counter c to the maximum potential, c = max(p(x, Y. z)).
Step 304. Look at the neighborhood of all the points where p(x, y, z) = c; at those points, I(x,y,z)=u~~~,(3(i,j)[~ I(x+i,y+j,z)+I(x,y,z)J, -_y=
where 1 if p(x + i, y + j, z) > p(x, y, z) 0 otherwise and a is a normalization constant such that uEE(3(i, j) = 1.
Step 306. Decrement c: c = c - 1.
Step 308. Determine whether c < 1. Repeat steps 304 and 306 until c < 1, at which time the process ends in step 310.
Once the illumination field is found, the illumination inhomogeneity in f(x, y, z) is corrected:
g(x~ Y~ z) = f (x~ Y~ z)k I (x, y, z) The new corrected image g(x, y, z) has smaller variations in the image intensity than the original image, particularly at those points at which the density should be constant.
In the real world the original image has been degraded in such a way that the discontinuity detection can be very hard. To avoid that problem, the image f(x, y, z) is pre-filtered with a non-linear structure preserving filter that reduced the noise around some of the streak artifacts. The filtered image was used for the estimation of the illumination profile.
Once the image has been corrected as described above, the streak artifact can be reduced. A 3 x 3 x 3 adaptive filter L(x, y, z) is used to reduce the streak artifact. The filter coefficient is a function if the propagation potential and the local image contrast and is given by L(x,y,z) _ (s r~) (s+(0.1+(p(xo +x, Yo '+' Y~zo +z)-P(xo~Yo~zo))Z)(g(xo '~' x~Yo +Y~zo ~' z)-g(xo~Yo~zo))Z) 2o where s is the noise estimation variance at point (x0, yo, zo) and r1 is the normalization constant.
Coefficients at neighboring pomis with different potential as well as those neighboring points whose density is very different will be very small. On the other hand, coefficients at neighboring points with similar potential and very similar density compared to the noise estimation variance are larger. This formulation reduces artifacts and preserves points where is a strong contrast between tissues.
The final processed image h(xo, yo, zo) at the point (xo, yo, zo) is given by h(xo ~ Yo ~ zo ) - ~ ~ ~ L(xo + x~ Yo + Y~ zo + z)g(xo + x~ Yo + Y~ zo + z) .
X-_y-_IZ-_I
The adaptive filtering process can be done several times so that it effectively removes most of the streaks present in the image.
1o In cases where the in-plane resolution is much finer than the slice thickness, the adaptive filter can be modified in such a way as to avoid degrading the image.
The image quality can also be improved by providing a space and streak oriented noise estimation. If that way, those regions are filtered where the streak artifacts are more important, and filtering can be avoided in those regions where the streaks are not so strong.
The algorithm has been tested on several CT images with a hip prosthesis. Fig.
shows a single slice of one of those images. The image slice contains the femur stem and the prosthesis cup. Fig. 4B shows the estimated potential field, while Fig. 4C
shows the estimation of the illumination field. Fig. 4D is the CT image after removing the illumination artifacts. Fig. 4E shows the filtered image after smoothing the artifacts with the adaptive filter. Figs. 5A and SB show the volume rendering of the hip from the raw data and the processed data, respectively. As one can see the volumetric rendering of the hip from the processed data allows to see the metal prosthesis as well as more detail in the bone that surrounds the prosthesis.
The embodiment disclosed above and other embodiments can be implemented in a system such as that shown in the block diagram of Fig. 6. The system 600 includes an input device 602 for input of the image data and the like. The information input through the input device 602 is received in the workstation 604, which has a storage device 606 such as a hard drive, a processing unit 608 for performing the processing disclosed above, and a graphics rendering engine 610 for preparing the final processed image for viewing, e.g., by surface rendering. An output device 612 can include a monitor for viewing the images rendered by the rendering engine 610, a further storage device such as a video recorder for recording the images, or both. Illustrative examples of the workstation 604 and the graphics rendering engine 610 are a Silicon Graphics Indigo workstation and an Irix Explorer 3D
graphics engine.
While a preferred embodiment has been set forth above in detail, those skilled in the art who have reviewed the present disclosure will readily appreciate that other embodiments can be realized within the scope of the invention. For example, numerical values, such as those given for a and 6, are illustrative rather than limiting. Also, the imaged object can be any human or animal body part or any other object. Moreover, the invention has applicability to imaging technologies other than CT. Therefore, the present invention should be construed as limited only by the appended claims.
to
Claims (20)
1. A method for imaging an object while reducing artifacts and improving a uniformity of illumination in the image, the method comprising:
(a) receiving raw imaging data from the object;
(b) estimating an illumination function from the raw imaging data;
(c) correcting the raw imaging data in accordance with the estimated illumination function to obtain an illumination corrected image;
(d) preparing an adaptive filter from the illumination corrected image; and (e) preparing a final processed image from the adaptive filter and the illumination corrected image.
(a) receiving raw imaging data from the object;
(b) estimating an illumination function from the raw imaging data;
(c) correcting the raw imaging data in accordance with the estimated illumination function to obtain an illumination corrected image;
(d) preparing an adaptive filter from the illumination corrected image; and (e) preparing a final processed image from the adaptive filter and the illumination corrected image.
2. The method of claim 1, wherein the step of receiving the raw imaging data is performed through computed tomography.
3. The method of claim 1, wherein the object is a body part.
4. The method of claim 1, wherein the step of estimating the illumination function comprises:
(i) creating a copy of the raw imaging data;
(ii) initializing a potential field to zero;
(iii) initializing a first counter to one;
(iv) gray scale dilating the raw imaging data in accordance with a value of the counter to obtain gray scale dilated data;
(v) setting the potential field equal to the first counter;
(vi) estimating a directional derivative of the illumination function;
(vii) reconstructing the illumination function from the directional derivative;
(viii) setting the raw imaging data equal to the gray scale dilated data;
(ix) incrementing the first counter; and (x) repeating steps (iv) through (ix) until no points in the potential field have changes in value.
(i) creating a copy of the raw imaging data;
(ii) initializing a potential field to zero;
(iii) initializing a first counter to one;
(iv) gray scale dilating the raw imaging data in accordance with a value of the counter to obtain gray scale dilated data;
(v) setting the potential field equal to the first counter;
(vi) estimating a directional derivative of the illumination function;
(vii) reconstructing the illumination function from the directional derivative;
(viii) setting the raw imaging data equal to the gray scale dilated data;
(ix) incrementing the first counter; and (x) repeating steps (iv) through (ix) until no points in the potential field have changes in value.
5. The method of claim 4, wherein the directional derivative at each point is estimated from the raw imaging data at that point and the raw imaging data at adjacent points.
6. The method of claim 5, wherein the illumination function is reconstructed by:
(A) initializing the illumination function to a constant equal to a mean image intensity;
(B) initializing a second counter to a maximum value of the potential field;
(C) recalculating the intensity function from the intensity function and the directional derivative of the intensity function;
(D) decrementing the second counter; and (E) repeating steps (C) and (D) until the second counter is below a threshold counter value.
(A) initializing the illumination function to a constant equal to a mean image intensity;
(B) initializing a second counter to a maximum value of the potential field;
(C) recalculating the intensity function from the intensity function and the directional derivative of the intensity function;
(D) decrementing the second counter; and (E) repeating steps (C) and (D) until the second counter is below a threshold counter value.
7. The method of claim 6, wherein the threshold counter value is one.
8. The method of claim 4, wherein the adaptive filter at each point is a function of the potential field and the illumination corrected image.
9. The method of claim 8, wherein the final processed image at each point is a function of the adaptive filter and the illumination corrected image at that point and at adjacent points.
10. The method of claim 1, wherein the final processed image at each point is a function of the adaptive filter and the illumination corrected image at that point and at adjacent points.
11. The method of claim 1, wherein step (e) comprises applying the adaptive filter to the illumination corrected image a plurality of times.
12. A system for imaging an object while reducing artifacts and improving a uniformity of illumination in the image, the system comprising:
an input for receiving raw imaging data;
a processor for estimating an illumination function from the raw imaging data, correcting the raw imaging data in accordance with the estimated illumination function to obtain an illumination corrected image, preparing an adaptive filter from the illumination corrected image, and preparing a final processed image from the adaptive filter and the illumination corrected image; and an output for outputting the final processed image.
an input for receiving raw imaging data;
a processor for estimating an illumination function from the raw imaging data, correcting the raw imaging data in accordance with the estimated illumination function to obtain an illumination corrected image, preparing an adaptive filter from the illumination corrected image, and preparing a final processed image from the adaptive filter and the illumination corrected image; and an output for outputting the final processed image.
13. The system of claim 12, wherein the processor estimates the illumination function by:
(i) creating a copy of the raw imaging data;
(ii) initializing a potential field to zero;
(iii) initializing a first counter to one;
(iv) gray scale dilating the raw imaging data in accordance with a value of the counter to obtain gray scale dilated data;
(v) setting the potential field equal to the first counter;
(vi) estimating a directional derivative of the illumination function;
(vii) reconstructing the illumination function from the directional derivative;
(viii) setting the raw imaging data equal to the gray scale dilated data;
(ix) incrementing the first counter; and (x) repeating steps (iv) through (ix) until no points in the potential field have changes in value.
(i) creating a copy of the raw imaging data;
(ii) initializing a potential field to zero;
(iii) initializing a first counter to one;
(iv) gray scale dilating the raw imaging data in accordance with a value of the counter to obtain gray scale dilated data;
(v) setting the potential field equal to the first counter;
(vi) estimating a directional derivative of the illumination function;
(vii) reconstructing the illumination function from the directional derivative;
(viii) setting the raw imaging data equal to the gray scale dilated data;
(ix) incrementing the first counter; and (x) repeating steps (iv) through (ix) until no points in the potential field have changes in value.
14. The system of claim 13, wherein the directional derivative at each point is estimated from the raw imaging data at that point and the raw imaging data at adjacent points.
15. The system of claim 14, wherein the illumination function is reconstructed by:
(A) initializing the illumination function to a constant equal to a mean image intensity;
(B) initializing a second counter to a maximum value of the potential field;
(C) recalculating the intensity function from the intensity function and the directional derivative of the intensity function;
(D) decrementing the second counter; and (E) repeating steps (C) and (D) until the second counter is below a threshold counter value.
(A) initializing the illumination function to a constant equal to a mean image intensity;
(B) initializing a second counter to a maximum value of the potential field;
(C) recalculating the intensity function from the intensity function and the directional derivative of the intensity function;
(D) decrementing the second counter; and (E) repeating steps (C) and (D) until the second counter is below a threshold counter value.
16. The system of claim 15, wherein the threshold counter value is one.
17. The system of claim 13, wherein the adaptive filter at each point is a function of the potential field and the illumination corrected image.
18. The system of claim 17, wherein the final processed image at each point is a function of the adaptive filter and the illumination corrected image at that point and at adjacent points.
19. The system of claim 12, wherein the final processed image at each point is a function of the adaptive filter and the illumination corrected image at that point and at adjacent points.
20. The system of claim 12, wherein the processor prepares the final processed image by applying the adaptive filter to the illumination corrected image a plurality of times.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US09/908,492 US6801646B1 (en) | 2001-07-19 | 2001-07-19 | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in CT imaging |
| US09/908,492 | 2001-07-19 | ||
| PCT/US2002/022707 WO2003009215A1 (en) | 2001-07-19 | 2002-07-18 | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in ct imaging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2452528A1 true CA2452528A1 (en) | 2003-01-30 |
Family
ID=25425893
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA 2452528 Abandoned CA2452528A1 (en) | 2001-07-19 | 2002-07-18 | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in ct imaging |
Country Status (5)
| Country | Link |
|---|---|
| US (2) | US6801646B1 (en) |
| EP (1) | EP1423819A1 (en) |
| JP (1) | JP2004535872A (en) |
| CA (1) | CA2452528A1 (en) |
| WO (1) | WO2003009215A1 (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6801646B1 (en) * | 2001-07-19 | 2004-10-05 | Virtualscopics, Llc | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in CT imaging |
| DE10238322A1 (en) * | 2002-08-21 | 2004-03-11 | Siemens Ag | Method for retrospective or window controlled filtering of computer tomography (CT) images e.g. for adapting sharpness and noise, involves automatic computation of CT-image sharpness of selected layer in primary data record |
| JP2005176988A (en) * | 2003-12-17 | 2005-07-07 | Ge Medical Systems Global Technology Co Llc | Data correcting method, and x-ray ct device |
| DE102004001273A1 (en) * | 2004-01-08 | 2005-08-04 | "Stiftung Caesar" (Center Of Advanced European Studies And Research) | Method for producing a sectional image |
| DE102004061507B4 (en) * | 2004-12-21 | 2007-04-12 | Siemens Ag | Method for correcting inhomogeneities in an image and imaging device therefor |
| JP5080986B2 (en) * | 2005-02-03 | 2012-11-21 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Diagnostic imaging system and diagnostic imaging method |
| US7453587B2 (en) * | 2005-12-07 | 2008-11-18 | Lexmark International, Inc. | Method for removing streaks from a scanned image |
| WO2008075272A1 (en) * | 2006-12-19 | 2008-06-26 | Koninklijke Philips Electronics N.V. | Apparatus and method for indicating likely computer-detected false positives in medical imaging data |
| US8582855B2 (en) * | 2007-01-04 | 2013-11-12 | Koninklijke Philips N.V. | Apparatus, method and computer program for producing a corrected image of a region of interest from acquired projection data |
| WO2009087777A1 (en) * | 2008-01-11 | 2009-07-16 | Shimadzu Corporation | Image processing method, its device and laminagraph device |
| JP5726288B2 (en) * | 2011-03-22 | 2015-05-27 | 株式会社日立メディコ | X-ray CT apparatus and method |
| GB201305755D0 (en) | 2013-03-28 | 2013-05-15 | Quanta Fluid Solutions Ltd | Re-Use of a Hemodialysis Cartridge |
| GB201409796D0 (en) | 2014-06-02 | 2014-07-16 | Quanta Fluid Solutions Ltd | Method of heat sanitization of a haemodialysis water circuit using a calculated dose |
| US9592020B2 (en) | 2014-06-23 | 2017-03-14 | Palodex Group Oy | System and method of artifact correction in 3D imaging |
| US10055671B2 (en) * | 2014-06-26 | 2018-08-21 | Siemens Aktiengesellschaft | Automatic assessment of perceptual visual quality of different image sets |
| GB201523104D0 (en) | 2015-12-30 | 2016-02-10 | Quanta Fluid Solutions Ltd | Dialysis machine |
| GB201622119D0 (en) | 2016-12-23 | 2017-02-08 | Quanta Dialysis Tech Ltd | Improved valve leak detection system |
| CN113520441B (en) * | 2021-08-03 | 2022-11-29 | 浙江大学 | Tissue imaging method and system for eliminating CT high-impedance artifact interference |
Family Cites Families (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4250387A (en) * | 1979-10-12 | 1981-02-10 | Emi Limited | Medical radiographic apparatus and method |
| US5305204A (en) * | 1989-07-19 | 1994-04-19 | Kabushiki Kaisha Toshiba | Digital image display apparatus with automatic window level and window width adjustment |
| US5225980A (en) * | 1991-11-27 | 1993-07-06 | General Electric Company | Reduction of image artifacts from support structures in tomographic imaging |
| US5311131A (en) * | 1992-05-15 | 1994-05-10 | Board Of Regents Of The University Of Washington | Magnetic resonance imaging using pattern recognition |
| US5561695A (en) | 1995-11-13 | 1996-10-01 | General Electric Company | Methods and apparatus for reducing image artifacts |
| JP3363735B2 (en) * | 1996-06-26 | 2003-01-08 | 松下電器産業株式会社 | X-ray imaging device |
| US5774521A (en) * | 1996-07-08 | 1998-06-30 | Cedars-Sinai Medical Center | Regularization technique for densitometric correction |
| IL119283A0 (en) * | 1996-09-19 | 1996-12-05 | Elscint Ltd | Adaptive filtering |
| US5883985A (en) * | 1996-12-10 | 1999-03-16 | General Electric Company | Method for compensating image data to adjust for characteristics of a network output device |
| US6246783B1 (en) * | 1997-09-17 | 2001-06-12 | General Electric Company | Iterative filter framework for medical images |
| US6125193A (en) * | 1998-06-01 | 2000-09-26 | Kabushiki Kaisha Toshiba | Method and system for high absorption object artifacts reduction and superposition |
| US6118845A (en) * | 1998-06-29 | 2000-09-12 | Surgical Navigation Technologies, Inc. | System and methods for the reduction and elimination of image artifacts in the calibration of X-ray imagers |
| US6169817B1 (en) * | 1998-11-04 | 2001-01-02 | University Of Rochester | System and method for 4D reconstruction and visualization |
| US6556720B1 (en) * | 1999-05-24 | 2003-04-29 | Ge Medical Systems Global Technology Company Llc | Method and apparatus for enhancing and correcting digital images |
| US6801646B1 (en) * | 2001-07-19 | 2004-10-05 | Virtualscopics, Llc | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in CT imaging |
| EP1649421B1 (en) * | 2003-07-18 | 2008-01-30 | Koninklijke Philips Electronics N.V. | Metal artifact correction in computed tomography |
-
2001
- 2001-07-19 US US09/908,492 patent/US6801646B1/en not_active Expired - Lifetime
-
2002
- 2002-07-18 EP EP02744872A patent/EP1423819A1/en not_active Withdrawn
- 2002-07-18 CA CA 2452528 patent/CA2452528A1/en not_active Abandoned
- 2002-07-18 JP JP2003514485A patent/JP2004535872A/en active Pending
- 2002-07-18 WO PCT/US2002/022707 patent/WO2003009215A1/en active Application Filing
-
2004
- 2004-08-18 US US10/920,298 patent/US7406211B2/en not_active Expired - Lifetime
Also Published As
| Publication number | Publication date |
|---|---|
| US7406211B2 (en) | 2008-07-29 |
| WO2003009215A1 (en) | 2003-01-30 |
| EP1423819A1 (en) | 2004-06-02 |
| JP2004535872A (en) | 2004-12-02 |
| US20050094857A1 (en) | 2005-05-05 |
| US6801646B1 (en) | 2004-10-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6801646B1 (en) | System and method for reducing or eliminating streak artifacts and illumination inhomogeneity in CT imaging | |
| Gjesteby et al. | Metal artifact reduction in CT: where are we after four decades? | |
| US10692189B2 (en) | Denoising medical images by learning sparse image representations with a deep unfolding approach using scan specific metadata | |
| US7873192B1 (en) | Method for providing a smoothed image | |
| Meilinger et al. | Metal artifact reduction in cone beam computed tomography using forward projected reconstruction information | |
| Liu et al. | Metal artifact reduction image reconstruction algorithm for CT of implanted metal orthopedic devices: a work in progress | |
| CN111656405A (en) | Reducing metal artifacts using deep learning | |
| CN111047524A (en) | Low-dose CT lung image denoising method based on deep convolutional neural network | |
| CN106530236B (en) | Medical image processing method and system | |
| Oulhaj et al. | Noise Reduction in Medical Images-comparison of noise removal algorithms | |
| Na'am et al. | Filter technique of medical image on multiple morphological gradient (MMG) method | |
| Luker et al. | Spiral CT of the temporal bone in unsedated pediatric patients. | |
| WO2008065394A1 (en) | Method and apparatus for reducing distortion in a computed tomography image | |
| Anhaus et al. | Nonlinearly scaled prior image‐controlled frequency split for high‐frequency metal artifact reduction in computed tomography | |
| Peng et al. | GPU-accelerated dynamic wavelet thresholding algorithm for X-ray CT metal artifact reduction | |
| CN112767265A (en) | Method, system and computer readable storage medium for correcting segmented image of removing metal artifact of CT image | |
| Saoji et al. | Speckle and rician noise removal from medical images and Ultrasound images | |
| Rodríguez-Gallo et al. | Inpainting-filtering for metal artifact reduction (IMIF-MAR) in computed tomography | |
| Moore | Medical image processing: the characterization of display changes using histogram entropy | |
| TWI593392B (en) | Metal Detection and Artifact Removal Methods in Computerized Tomography Images | |
| CN108171768A (en) | A kind of low dosage CBCT image rebuilding methods based on BM3D | |
| Qin et al. | Metal artifact reduction in kVCT images via L1 sparse regularization of MVCBCT prior images | |
| Anhaus et al. | A New Iterative Metal Artifact Reduction Algorithm for Both Energy-Integrating and Photon-Counting CT Systems | |
| Zhang et al. | Denoising of degenerative lumbar spine lesions MRI images using block-matching and 3D filtering | |
| Setty et al. | Details enhancement of MRI image using illumination and reflectance estimation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| FZDE | Discontinued |