Method for inhibiting metal artifacts

Abstract

The invention discloses a method for inhibiting metal artifacts, which comprises the following steps of: (1) acquiring a CBCT (Cone Beam Computed Tomography) reconstructed image as an initial reconstructed image, carrying out front projection on the initial reconstructed image to obtain a projected image, and carrying out back projection reconstruction on the projected image to obtain a reconstructed image; (2) repeating the step (1) for at least one time to respectively obtain a second projection image and a second reconstruction image obtained at the last time and a first projection image and a first reconstruction image obtained at the previous time; (3) obtaining a difference image between the first reconstructed image and the second reconstructed image obtained in the step (2); and (4) making a difference between the first reconstructed image and the difference image obtained in the step (3), and taking a result as a final image for suppressing the metal artifact. According to the invention, the generation of secondary artifacts can be greatly avoided while the metal artifacts are inhibited.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method for suppressing metal artifacts. Background Art

[0002] Cone-Beam Computed Tomography (CBCT) has been widely used in the fields of orthodontics, orthopedic surgery, radiotherapy, etc. in recent years due to its advantages such as fast acquisition speed, low radiation dose and high pixel accuracy. However, if there are high attenuation coefficient materials (typically metals), it will cause serious metal artifacts in the CBCT reconstructed images. The specific characteristics are scattered light and dark stripes around the high attenuation coefficient objects, large-area light and dark stripes, etc., which will cause great damage to the image quality. Therefore, it is of great significance to suppress and eliminate metal artifacts.

[0003] Most current methods for suppressing metal artifacts rely on segmentation. For example, the metal in the reconstructed image is spatially segmented, and then the area containing the metal in the projection image is set to 0 and then interpolated and backfilled through the surrounding pixel values. This type of method is suitable for situations where the metal structure is simple and small in size, but when the metal structure is complex or placed in a complex anatomical environment, the metal segmentation method will produce additional artifacts (secondary artifacts) due to the inability to accurately segment the metal area and the difficulty in accurately interpolating the corresponding projection area, which greatly interferes with the diagnosis. Summary of the invention

[0004] Purpose of the invention: In view of the above-mentioned deficiencies, the present invention provides a method for suppressing metal artifacts with a new idea, which does not rely on the segmentation of specific areas. While suppressing metal artifacts, it can greatly avoid the generation of secondary artifacts.

[0005] Technical solution:

[0006] The present invention provides a method for suppressing metal artifacts, comprising the steps of:

[0007] (1) obtaining a CBCT reconstructed image as an initial reconstructed image, forward-projecting the image to obtain a projection image, and back-projecting the projection image to obtain a reconstructed image;

[0008] (2) repeating step (1) at least once to obtain the second projection image and the second reconstructed image obtained last time and the first projection image and the first reconstructed image obtained last time respectively;

[0009] (3) obtaining a difference image between the first reconstructed image obtained in step (2) and the second reconstructed image;

[0010] (4) Subtract the first reconstructed image from the difference image obtained in step (3), and use the result as the final image with metal artifacts suppressed.

[0011] Specifically, after the first reconstructed image and the second reconstructed image are obtained in step (2), they are corrected respectively, and the difference between the voxel values ​​of the areas after removing the metal area and its artifact area in the two corrected images and the voxel values ​​of the corresponding areas in the initial reconstructed image is within a set range.

[0012] More specifically, the difference between the voxel values ​​of the area after removing the metal area and its artifact area in the two images and the voxel values ​​of the corresponding area in the initial reconstructed image is set to the difference in average voxel values ​​of the area after removing the metal area and its artifact area in the two images.

[0013] Furthermore, the correction is specifically:

[0014] Calculate the correction coefficients c1 and c2 of the first reconstructed image and the second reconstructed image respectively, and multiply the voxel value of each voxel in the first reconstructed image and the second reconstructed image by its corresponding correction coefficient, so as to complete the correction of the first reconstructed image and the second reconstructed image;

[0015] The correction coefficients c1 and c2 of the first reconstructed image and the second reconstructed image are calculated as follows:

[0016]

[0017] Among them, x im0 (i), x im1 (i), x im2 (i) are the voxel values ​​of the i-th voxel in the initial reconstructed image, the first reconstructed image, and the second reconstructed image, respectively; N represents the number of voxels in the initial reconstructed image, the first reconstructed image, and the second reconstructed image.

[0018] Specifically, in step (3), a difference image between the corrected first reconstructed image and the second reconstructed image is obtained.

[0019] More specifically, the difference image between two images is obtained by subtracting the two images.

[0020] More specifically, after obtaining the difference image, it is smoothed to obtain a smoothed difference image. In the step (4), the first reconstructed image or the corrected first reconstructed image is subtracted from the smoothed difference image, and the result is used as the final image suppressing metal artifacts.

[0021] Specifically, in step (2), step (1) is repeated once to obtain a second reconstructed image, and the reconstructed image obtained in step (1) is used as the first reconstructed image.

[0022] Beneficial effect: According to the present invention, in the images before and after multiple front projections and back projections reconstruction, the changes in the metal area and the area of ​​artifacts near it are different from those in other areas of the image. Therefore, the metal area and the area of ​​artifacts near it in the image are extracted, so that the metal artifacts can be captured and artifacts can be suppressed accordingly, thereby eliminating the need to rely on the segmentation of specific areas. While suppressing the metal artifacts, the generation of secondary artifacts is greatly avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are merely embodiments of the present invention, and ordinary technicians in this field can obtain other drawings based on these drawings without creative work.

[0024] Figure 1 This is an example flow chart of the method for suppressing metal artifacts of the present invention;

[0025] Figure 2 These are example diagrams of the effect of suppressing metal artifacts using the method of the present invention, wherein the left figure is a reconstructed image without suppressing metal artifacts, and the right figure is a reconstructed image processed by the method of suppressing metal artifacts of the present invention. It can be observed in the yellow frame that the metal peripheral scattering artifacts are partially suppressed. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present application is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0027] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. "Include" or "comprising" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

[0028] The method for suppressing metal artifacts of the present invention is as follows Figure 1 As shown, the steps include:

[0029] S1, obtaining a CBCT reconstructed image as an initial reconstructed image im0, forward-projecting it to obtain a projection image, and back-projecting and reconstructing the projection image to obtain a reconstructed image;

[0030] In the present invention, an image is acquired through CBCT, and then reconstructed through a reconstruction algorithm to obtain a CBCT reconstructed image, wherein the reconstruction algorithm may adopt the FDK (Feldkamp-Davis-Kress) reconstruction algorithm.

[0031] In the present invention, the front projection may adopt the FP (Forward-Projection) algorithm.

[0032] In the present invention, the back-projection reconstruction may adopt the BP (Back-Projection) algorithm.

[0033] In the present invention, after the reconstruction by forward projection and back projection, the reconstructed image obtained has a smaller change in the voxel value of the area far from the metal area than the initial reconstructed image im0, and this difference in change distinguishes the metal area and the area of ​​the nearby artifacts from other areas. Therefore, through multiple forward projection and back projection reconstruction operations, in the images before and after the operation, there is a difference in the change of the metal area and the area of ​​the nearby artifacts from other areas in the image, and the metal area and the area of ​​the nearby artifacts in the image can be extracted accordingly.

[0034] S2, repeat S1 at least once, to obtain the second projection image p2 and the second reconstructed image im2 obtained last time and the first projection image p1 and the first reconstructed image im1 obtained last time respectively;

[0035] In the present invention, S1 can be repeated multiple times so that after multiple front projection and back projection reconstruction operations, the difference between the changes in the metal area and the area of ​​its nearby artifacts in the image and the changes in other areas in the image is greater. Therefore, the second reconstructed image im2 and the first reconstructed image im1 both contain metal artifacts with the same characteristics but different degrees, so that the metal area and the area of ​​its nearby artifacts can be extracted.

[0036] Preferably, in this embodiment, S1 is repeated once, that is, after two front projection and back projection reconstruction operations, the first projection image p1 and the first reconstructed image im1 are obtained, and the second projection image p2 and the second reconstructed image im2 are obtained for the second time, wherein, compared with the first reconstructed image im1, the second reconstructed image im2 has a larger change in the metal area (usually also including the nearby artifact area) than in other areas in the image. In fact, the metal artifact is the error about the metal introduced in the three-dimensional reconstruction process. Therefore, the present application actively performs multiple three-dimensional reconstructions to introduce new errors associated with the metal, so that the change near the metal area is more obvious than the change far away from the metal area, so that the metal area and the area of ​​the artifact near it can be extracted.

[0037] S3, correcting the first reconstructed image im1 and the second reconstructed image im2 obtained in S2, so that the voxel values ​​far from the metal area are corrected to be consistent with the voxel values ​​far from the metal area in the initial reconstructed image im0. In the present invention, the far from metal area is the area after removing the aforementioned metal area and its artifact area from the image;

[0038] In the present invention, the voxel values ​​of the two images far away from the metal area are consistent, that is, the difference in the voxel values ​​of the two images far away from the metal area is within a set range. It can generally be set to that the difference in the average voxel values ​​of the area after removing the metal area and its artifact area in the two images is within the set range. Furthermore, the consistency of the voxel values ​​of the area in the two images can be set to that the average voxel values ​​of the area in the two images are equal.

[0039] Due to factors such as truncation, multiple forward projection and back-projection reconstructions will lead to global deviations in the voxel values ​​in the reconstructed image. If the first reconstructed image im1 is directly subtracted from the second reconstructed image im2, in addition to the differences in the metal region and the region of the nearby artifacts, the global deviation of the voxel values ​​will also be captured. Therefore, in order to highlight the changes in the metal region and the region of the nearby artifacts caused by the forward projection and back-projection reconstruction, it is necessary to make an overall correction to the first reconstructed image im1 before subtracting it from the second reconstructed image im2, so that the voxel values ​​far from the metal region in the two images after correction are almost the same.

[0040] Specifically, the voxel values ​​far from the metal region in the first reconstructed image im1 and the second reconstructed image im2 may be corrected to be consistent with the voxel values ​​far from the metal region in the initial reconstructed image im0 .

[0041] Assuming that the sum of the voxel values ​​of all voxels in the corrected reconstructed image is approximately equal, therefore, using this equal relationship, the correction coefficients c1 and c2 of the first reconstructed image im1 and the second reconstructed image im2 are calculated respectively, and the voxel value of each voxel in the first reconstructed image im1 and the second reconstructed image im2 is multiplied by its corresponding correction coefficient to complete the correction of the first reconstructed image im1 and the second reconstructed image im2.

[0042] The correction coefficients c1 and c2 of the first reconstructed image im1 and the second reconstructed image im2 are calculated as follows:

[0043]

[0044]

[0045] Among them, x im0 (i), x im1 (i), x im2(i) is the voxel value of the i-th voxel in the initial reconstructed image im0, the first reconstructed image im1, and the second reconstructed image im2, respectively; N is the number of voxels in the initial reconstructed image im0, the first reconstructed image im1, and the second reconstructed image im2.

[0046] In the present invention, the difference in voxel values ​​in the two images far from the metal region may also be set as the difference in the variance, mean square error, etc. of the voxel values ​​in the region in the two images being within a set range.

[0047] S4, obtaining a difference image im3 between the first reconstructed image im1 and the second reconstructed image im2;

[0048] Specifically, the first reconstructed image im1 corrected by S3 is subtracted from the second reconstructed image im2 to obtain a difference image im3;

[0049] In the present invention, the first reconstructed image obtained by S2 can be directly subtracted from the second reconstructed image to obtain a difference image.

[0050] In the present invention, the first reconstructed image im1 corrected by S3 is subtracted from the second reconstructed image im2. Specifically, the voxel values ​​of the corresponding voxels in the first reconstructed image im1 corrected by S3 and the second reconstructed image im2 are subtracted and used as the voxel values ​​of the corresponding voxels in the difference image im3, thereby obtaining the difference image im3. The difference image im3 is the extracted metal area and the area of ​​its nearby artifacts, that is, the extracted image containing metal and its artifacts.

[0051] In the present invention, since S3 corrects the voxel values ​​of the first reconstructed image im1 and the second reconstructed image im2 away from the metal area to be consistent with the voxel values ​​of the initial reconstructed image im0 away from the metal area, the difference image im3 is an image containing metal and its artifacts, but the metal artifacts in the image are not absolutely accurate in space and value. There is a certain degree of positional deviation between the difference image im3 and the ideal metal area and its artifact area, resulting in poor effect after removal. Therefore, the present invention also needs to smooth the difference image im3 to correct the spatial and numerical deviation of the metal artifacts to obtain a smoothed difference image im4. The smoothing process can be performed in any existing manner. In this example, a three-dimensional Gaussian filter can be used, wherein the filter kernel size is 5×5.

[0052] S5. Subtract the first reconstructed image im1 corrected by S3 from the smoothed difference image im4 obtained by S4, and use the result as the final image with suppressed metal artifacts.

[0053] In the present invention, the first reconstructed image im1 corrected by S3 and the difference image im3 obtained by S4 may be subtracted, and the result may be used as the final image for suppressing metal artifacts.

[0054] In the present invention, metal artifacts are introduced in the reconstruction calculation process, so the metal artifacts in the first reconstructed image im1 are more prominent than those in the initial reconstructed image im0; the metal artifacts in the second reconstructed image im2 are more prominent than those in the first reconstructed image im1. This "prominence" is not necessarily at the visual perception level, but more at the computational feature level, which can be understood as a kind of adjustment or domestication of the artifacts, making it more suitable for unified processing. At the same time, assuming that there is no metal, multiple forward / back projections themselves will also lead to the introduction of computational errors. This error has nothing to do with whether there is metal and is relatively small. In addition, the overall image quality of the second reconstructed image im2 is slightly lower than that of the first reconstructed image im1, and the metal artifacts on the second reconstructed image im2 are too obvious. After subtracting from the smooth difference image im4, metal artifacts may still exist. Therefore, while considering suppressing metal artifacts, the present invention selects the first reconstructed image im1 and the smooth difference image im4 obtained by S4 for subtraction, and uses the result as the image that finally suppresses metal artifacts.

[0055] In the present invention, compared with the result not processed by the metal artifact removal algorithm, the metal artifacts in the result processed by this example are weakened, such as Figure 2 shown.

[0056] The present invention can suppress artifacts by directly or indirectly capturing metal artifacts without relying on the segmentation of specific areas. Therefore, while suppressing metal artifacts, the generation of secondary artifacts is greatly avoided. The method of suppressing metal artifacts of the present invention is particularly suitable for complex metal structures or when they are placed in complex anatomical environments. Since it does not rely on the segmentation of specific areas and does not require corresponding interpolation operations, it can ensure the basic effect of removing metal artifacts without introducing secondary artifacts.

[0057] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0058] The embodiments of the present invention are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for suppressing metal artifacts, characterized in that: Includes steps: (1) obtaining a CBCT reconstructed image as an initial reconstructed image, forward-projecting the image to obtain a projection image, and back-projecting the projection image to obtain a reconstructed image; (2) repeating step (1) at least once to obtain the second projection image and the second reconstructed image obtained last time and the first projection image and the first reconstructed image obtained last time respectively; (3) obtaining a difference image between the first reconstructed image obtained in step (2) and the second reconstructed image; (4) Subtract the first reconstructed image from the difference image obtained in step (3), and use the result as the final image with metal artifacts suppressed.

2. The method for suppressing metal artifacts according to claim 1, characterized in that: After the first reconstructed image and the second reconstructed image are obtained in step (2), they are corrected respectively, and the difference between the voxel values ​​of the areas after removing the metal area and its artifact area in the two corrected images and the voxel values ​​of the corresponding areas in the initial reconstructed image is within a set range.

3. The method for suppressing metal artifacts according to claim 2, characterized in that: The difference between the voxel value of the area after removing the metal area and its artifact area in the two images and the voxel value of the corresponding area in the initial reconstructed image is set to the difference in average voxel value of the area after removing the metal area and its artifact area in the two images.

4. The method for suppressing metal artifacts according to claim 3, characterized in that: The amendments are specifically: Calculate the correction coefficients c1 and c2 of the first reconstructed image and the second reconstructed image respectively, and multiply the voxel value of each voxel in the first reconstructed image and the second reconstructed image by its corresponding correction coefficient, so as to complete the correction of the first reconstructed image and the second reconstructed image; The correction coefficients c1 and c2 of the first reconstructed image and the second reconstructed image are calculated as follows: Among them, x im0 (i), x im1 (i), x im2 (i) are the voxel values ​​of the i-th voxel in the initial reconstructed image, the first reconstructed image, and the second reconstructed image, respectively; N represents the number of voxels in the initial reconstructed image, the first reconstructed image, and the second reconstructed image.

5. The method for suppressing metal artifacts according to claim 2, characterized in that: In the step (3), a difference image between the corrected first reconstructed image and the second reconstructed image is obtained.

6. The method for suppressing metal artifacts according to claim 1 or 5, characterized in that: The difference image between two images is obtained by subtracting the two images.

7. The method for suppressing metal artifacts according to claim 1 or 5, characterized in that: After obtaining the difference image, it is smoothed to obtain a smoothed difference image. In the step (4), the first reconstructed image or the corrected first reconstructed image is subtracted from the smoothed difference image, and the result is used as the final image suppressing metal artifacts.

8. The method for suppressing metal artifacts according to claim 1, characterized in that: In the step (2), step (1) is repeated once to obtain a second reconstructed image, and the reconstructed image obtained in step (1) is used as the first reconstructed image.

Images