Medical image processing method, apparatus, and medical imaging device

Abstract

The application discloses a medical image processing method and device and medical imaging equipment. When there is a ring artifact in an initial medical image sequence, the radius of the ring artifact in the initial medical image sequence is acquired, the artifact slice image sequence can be determined in a plurality of slice images included in the initial medical image sequence according to the radius of the ring artifact, the ring artifact is eliminated from a target slice image included in the artifact slice image sequence, and a target medical image sequence without the ring artifact is obtained. Through automatic identification of the artifact slice image sequence in the plurality of slice images included in the initial medical image sequence and elimination of the ring artifact from the artifact slice image sequence, automation of slice image ring artifact elimination is realized, and image quality and authenticity are improved. Further, through automatic identification of the artifact slice image sequence according to the radius of the ring artifact, the removal efficiency of the ring artifact can be improved.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, and in particular to a medical image processing method, apparatus, and medical imaging equipment. Background Technology

[0002] Ring artifacts are a common type of artifact in cone-beam computed tomography (CBCT), primarily caused by inconsistent detector responses. The presence of ring artifacts degrades image quality, complicating subsequent diagnosis and analysis.

[0003] In related technologies, ring artifacts caused by inconsistent detector response can be eliminated by allowing the detector to remain stationary for an extended period to allow it to return to normal before re-acquiring CBCT images. However, the method of stationary detector placement needs improvement. Summary of the Invention

[0004] The embodiments described in this specification aim to at least partially solve one of the technical problems in the related art. To this end, the embodiments of this specification propose a medical image processing method, apparatus, and medical imaging device.

[0005] This specification provides a medical image processing method, the method comprising:

[0006] Obtain the radius of the annular artifact in the initial medical image sequence; wherein the initial medical image sequence includes multiple slice images;

[0007] Based on the radius of the annular artifact, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence; wherein, the artifact slice image sequence includes at least one target slice image containing the annular artifact;

[0008] The annular artifact is removed from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

[0009] In one embodiment, the initial medical image sequence corresponds to reconstructed imaging parameters and source location parameters. The reconstructed imaging parameters represent the parameters used to reconstruct the initial medical image sequence; the source location parameters represent the location of the radiation source; determining the artifact slice image sequence from the multiple slice images included in the initial medical image sequence based on the radius of the annular artifact includes:

[0010] Based on the radius of the annular artifact, the reconstructed imaging parameters, and the source location parameters, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence.

[0011] In one embodiment, the reconstructed imaging parameters include the total number of reconstructed slices, pixel spacing, and slice layer thickness; determining the artifact slice image sequence from the multiple slice images included in the initial medical image sequence based on the radius of the annular artifact and the reconstructed imaging parameters includes:

[0012] The imaging field of view that causes the annular artifact is determined based on the radius of the annular artifact;

[0013] The starting slice number and ending slice number where the annular artifact exists are determined based on the imaging field size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness.

[0014] The slice images located between the starting slice number and the ending slice number are used as the artifact slice image sequence;

[0015] The starting slice number and the ending slice number are determined using the following formulas:

[0016]

[0017]

[0018] Where zs is the starting slice number, ze is the ending slice number, SlicesFrames is the total number of slice images in the initial medical image sequence, F is the imaging field of view size, pixelpitch is the pixel pitch, SLicesThickness is the slice thickness, SOD is the vertical distance between the X-ray source and the rotation center of the imaging device, and SID is the vertical distance between the X-ray source and the center of the detector.

[0019] In one implementation, obtaining the radius of the annular artifact in the initial medical image sequence includes:

[0020] At least some slice images in the initial medical image sequence are converted to polar coordinates to obtain a polar coordinate image sequence.

[0021] Based on the polar coordinate system image sequence, a first gray-level distribution curve is obtained; wherein, the first gray-level distribution curve represents the average pixel value of the at least part of the slice image corresponding to different radii in the polar coordinate system;

[0022] Extract the gray-level distortion points from the first gray-level distribution curve, and use the radius corresponding to the gray-level distortion points as the radius of the annular artifact.

[0023] In one implementation, obtaining a first grayscale distribution curve based on the polar coordinate system image sequence includes:

[0024] Based on the number of pixels in polar coordinates, the average pixel data of each slice image in the polar coordinate system image sequence is obtained by averaging the pixels in the angular direction in any polar coordinate system image in the polar coordinate system image sequence at different radii.

[0025] The first grayscale distribution curve is obtained by averaging the mean pixel data corresponding to different radii of each slice image along the slice direction.

[0026] In one implementation, extracting gray-level distortion points from the first gray-level distribution curve includes:

[0027] The first grayscale distribution curve is smoothed to obtain the second grayscale distribution curve;

[0028] The target grayscale distribution curve is obtained by subtracting the first grayscale distribution curve from the first grayscale distribution curve.

[0029] The peak of the target grayscale distribution curve is taken as the grayscale distortion point.

[0030] In one embodiment, the step of removing the annular artifact from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence includes:

[0031] The initial artifact enhancement image is obtained by performing pixel-by-pixel averaging on the target slice images included in the artifact slice image sequence; wherein, the initial artifact enhancement image is located in the Cartesian coordinate system;

[0032] The initial artifact enhancement image is transformed to polar coordinates to obtain the initial linear artifact image;

[0033] The target line artifact image is obtained by filtering the initial line artifact image.

[0034] The target linear artifact image is transformed into a Cartesian coordinate system to obtain an enhanced target artifact image;

[0035] Based on the target artifact enhancement image, artifact removal is performed on the target slice images included in the artifact slice image sequence to obtain the target medical image sequence.

[0036] In one implementation, the step of removing artifacts from the target slice images included in the artifact slice image sequence based on the target artifact enhancement image to obtain a target medical image sequence includes:

[0037] The corresponding pixel subtraction operation is performed between any target slice image in the artifact slice image sequence and the target artifact enhancement image to obtain the artifact-removed slice image corresponding to the target slice image; the target medical image sequence is obtained by replacing the target slice image with the artifact-removed slice image.

[0038] In one implementation, the step of removing artifacts from the target slice images included in the artifact slice image sequence based on the target artifact enhancement image to obtain a target medical image sequence includes:

[0039] Based on the artifact adjustment data and the target artifact enhancement image, artifact removal pixel data is determined. A corresponding pixel subtraction operation is performed between the artifact slice image sequence and the artifact removal pixel data to obtain the artifact removal slice image corresponding to any target slice image. The target medical image sequence is obtained by replacing any target slice image with the artifact removal slice image.

[0040] This specification provides a medical image processing apparatus, the apparatus comprising:

[0041] An initial sequence acquisition module is used to acquire the radius of the annular artifact in the initial medical image sequence; wherein, the annular artifact in the initial medical image sequence includes multiple slice images;

[0042] The artifact sequence determination module is used to determine an artifact slice image sequence from multiple slice images included in the initial medical image sequence based on the radius of the annular artifact; wherein the artifact slice image sequence includes at least one target slice image containing the annular artifact;

[0043] The ring artifact removal module is used to remove the ring artifact from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

[0044] This specification provides a medical imaging device, which includes: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, which, when executed by the one or more processors, cause the one or more processors to perform the steps of the method described in any of the above embodiments.

[0045] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the above embodiments.

[0046] This specification provides a computer program product that includes instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0047] In the above-described embodiment, when annular artifacts exist in the initial medical image sequence, the radius of the annular artifacts in the initial medical image sequence is obtained. Based on the radius of the annular artifacts, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence. The annular artifacts are then eliminated from the artifact slice image sequence to obtain a target medical image sequence free of annular artifacts. This embodiment does not employ a long-term static detector placement method but proposes a novel image post-processing approach. By automatically identifying the artifact slice image sequence from the multiple slice images included in the initial medical image sequence and eliminating annular artifacts from the artifact slice image sequence, the elimination of annular artifacts in slice images is automated, improving image quality and realism. Furthermore, by automatically identifying the artifact slice image sequence based on the radius of the annular artifacts, the removal efficiency of annular artifacts can be improved. Attached Figure Description

[0048] Figure 1a A schematic diagram of the detector distribution provided for embodiments of this specification;

[0049] Figure 1b A schematic flowchart illustrating the medical image processing method provided in the embodiments of this specification;

[0050] Figure 2a A flowchart illustrating the process of determining a target slice image provided for embodiments of this specification;

[0051] Figure 2b A schematic diagram illustrating the imaging field of view dimensions provided for embodiments of this specification;

[0052] Figure 3 A schematic diagram showing the termination slice number and the start slice number with ring artifacts provided for the embodiments of this specification;

[0053] Figure 4a A flowchart illustrating the determination of the radius of a ring artifact provided for embodiments of this specification;

[0054] Figure 4b A schematic diagram of a medical image in the Cartesian coordinate system provided for embodiments of this specification;

[0055] Figure 4c A schematic diagram of a medical image in polar coordinates provided for an embodiment of this specification;

[0056] Figure 4dA flowchart illustrating the determination of the radius of a ring artifact provided for embodiments of this specification;

[0057] Figure 5a A flowchart illustrating the determination of the radius of a ring artifact provided for embodiments of this specification;

[0058] Figure 5b A schematic diagram of the mean pixel data corresponding to the polar coordinate system image sequence provided in the embodiments of this specification;

[0059] Figure 5c A schematic diagram of a polar coordinate system image provided for embodiments of this specification;

[0060] Figure 5d A schematic diagram of the mean pixel data corresponding to the polar coordinate system image sequence provided in the embodiments of this specification;

[0061] Figure 5e A schematic diagram of the first grayscale distribution curve provided for the embodiments of this specification;

[0062] Figure 5f A flowchart illustrating the determination of the first grayscale distribution curve provided for the implementation of this specification;

[0063] Figure 6 A flowchart illustrating the determination of the radius of a ring artifact provided for embodiments of this specification;

[0064] Figure 7a A schematic diagram illustrating the process of obtaining the target medical image sequence provided for the embodiments of this specification;

[0065] Figure 7b A schematic diagram of the target slice image provided for the embodiments of this specification;

[0066] Figure 8 A schematic flowchart illustrating the medical image processing method provided in the embodiments of this specification;

[0067] Figure 9 A schematic diagram of a medical image processing apparatus provided for embodiments of this specification. Detailed Implementation

[0068] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0069] Cone-beam CT uses a cone-beam X-ray source and a flat panel detector to acquire projection data of the object being measured and reconstruct a continuous sequence of slice images. It can apply three-dimensional tomographic imaging technology to the digital subtraction angiography (DSA) flat panel detector system, giving DSA equipment higher soft tissue resolution.

[0070] Low-density ring artifacts are a type of artifact in cone-beam CT, primarily caused by inconsistent response across different detector regions. For example, before acquiring projection data of the object being measured, the cone-beam CT detector may continuously expose within a small field of view, while other areas outside this small field of view remain unexposed. Consequently, the detector state in this small field of view changes, differing from that in other areas. When the cone-beam CT acquires the object, the uneven response between the small field of view and other detector regions results in low-density ring artifacts, appearing as dark circles, in the reconstructed sequence of slices. The presence of low-density ring artifacts degrades medical image quality, complicating subsequent diagnosis and analysis.

[0071] CBCT (Cone-Beam Computed Tomography) is a medical imaging technique that offers both small and large field-of-view imaging. During CBCT imaging, 2D images are acquired, and this acquisition involves changes in the field of view. For example, in interventional treatment, a small field-of-view scan is needed to magnify the lesion area, revealing its minute structures and providing more detailed information. In this case, the detectors in the small field-of-view area are continuously exposed to X-rays, causing different impacts on the detectors in that area compared to those outside the small field-of-view, resulting in inconsistent detector states. After the interventional treatment, a large field-of-view scan is used to assess the overall treatment effect. However, overlapping and non-overlapping areas exist between the detectors in the small and large field-of-view areas. Inconsistent detector states in the overlapping and non-overlapping areas lead to inconsistent responses and low-density ring artifacts. For example, please refer to [link to relevant documentation]. Figure 1aAfter prolonged surgical treatment of a lesion within a small 20*20 field of view, it is necessary to switch to a larger 30*30 field of view to allow the patient to observe the surrounding area and treatment effectiveness. During the surgical procedure, the detector's response area is the central 20*20 area A; after the procedure, the detector's response area is the entire detector area, i.e., the 30*30 area B. The area between area A and area B is designated as area C. Area B includes both areas A and C, and the detectors within area A and area C are not identical. When examining the overall lesion area after surgery, imaging within the larger field of view corresponding to area B reveals inconsistent responses between the detectors in area A and area C. Therefore, low-density ring artifacts are present in the medical images obtained under a 30*30 field of view.

[0072] In related technologies, low-density ring artifacts caused by detector inconsistencies can be eliminated by allowing the detector to return to normal operation for an extended period and then re-acquiring CBCT image signals. Alternatively, low-density ring artifacts caused by detector inconsistencies can be eliminated by calibrating the detector and performing high-dose exposure.

[0073] However, prolonged static placement of the detector can disrupt the continuity of clinical diagnosis. High-dose exposure poses radiation risks to patients, especially pregnant women, children, and the elderly, causing greater risks and discomfort. Therefore, prolonged static placement of the detector and high-dose exposure are not suitable for clinical settings.

[0074] In some implementations, the presence of circles can be detected layer by layer on each slice image using Hough transform to determine the presence of low-density annular artifacts. Alternatively, each slice image can be converted to polar coordinates before applying Hough transform layer by layer to detect the presence of straight lines, thus determining the presence of low-density annular artifacts. However, the HU values ​​of soft tissue regions and annular artifacts in cone-beam CT images are close, and the detector response is not uniform across the entire image. Furthermore, the intensity of annular artifacts is not entirely consistent, leading to potential false negatives or missed detections when using Hough transform to detect circles or straight lines on slice images. Additionally, layer-by-layer detection is resource-intensive and time-consuming.

[0075] Based on this, this specification provides a medical image processing method. When annular artifacts exist in the initial medical image sequence, the radius of the annular artifacts in the initial medical image sequence is obtained. Based on the radius of the annular artifacts, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence. The annular artifacts are then eliminated from the artifact slice image sequence to obtain a target medical image sequence free of annular artifacts. This embodiment does not employ a long-term static detector approach but proposes a novel image post-processing method. By automatically identifying the artifact slice image sequence from the multiple slice images included in the initial medical image sequence and eliminating the annular artifacts from the artifact slice image sequence, the elimination of annular artifacts in slice images is automated, improving image quality and realism. Furthermore, by automatically identifying the artifact slice image sequence based on the radius of the annular artifacts, the removal efficiency of annular artifacts can be improved.

[0076] The method described in this specification can be used in the following application scenarios: During a patient's surgery, cone-beam CT can be used to acquire images of the lesion area to determine the condition of the lesion. Cone-beam CT can acquire medical images of the lesion area from different angles and positions, obtaining multiple two-dimensional medical images. These multiple two-dimensional medical images are then reconstructed into a three-dimensional medical image sequence using a reconstruction algorithm. The slice images included in the medical image sequence are converted to polar coordinates to obtain a polar coordinate image sequence. Based on the polar coordinate image sequence, the radius of the low-density ring artifact can be determined, and thus the imaging field of view size can be determined based on the radius of the low-density ring artifact. Further, based on the imaging field of view size, the total number of reconstructed slices, pixel spacing, slice thickness, and geometric relationships related to the X-ray source position, the starting slice number and ending slice number where the low-density ring artifact exists can be determined. Finally, a correction algorithm is used to eliminate the ring artifact in the slice images between the starting and ending slice numbers to obtain the target medical image sequence.

[0077] It should be noted that the annular artifacts described in this specification are different from common annular artifacts; they are low-density annular artifacts. Low-density annular artifacts are usually caused by specific problems in the detector or optical system, such as inconsistent response in different areas of the detector. Common annular artifacts, on the other hand, are caused by scattering and absorption phenomena resulting from the X-ray beam passing through the uneven tissue density within the patient's body during a CT scan. Although low-density annular artifacts can interfere with the doctor's interpretation of CT images, the grayscale value of the annular region corresponding to the low-density annular artifact will be slightly darker or blacker than the surrounding area. This is because the pixel values ​​in this region have been over-smoothed, resulting in a lower grayscale value.

[0078] This specification provides a medical image processing method. Please refer to [link / reference]. Figure 1bThe medical image processing method may include the following steps:

[0079] S110. Obtain the radius of the ring artifact in the initial medical image sequence.

[0080] The initial medical image sequence contains ring artifacts and comprises multiple slice images. The initial medical image sequence is a collection of medical images, typically consisting of multiple slices or layers, each acquired from different angles and orientations. Ring artifacts are primarily caused by inconsistent detector responses across different regions, and they negatively impact the quality and readability of the medical images.

[0081] Specifically, medical equipment can be used to examine patients and acquire multiple two-dimensional projected medical images of the patient's target area from different angles. Then, a reconstruction algorithm can be used to reconstruct a three-dimensional model from this series of two-dimensional projected medical images, resulting in an initial medical image sequence. When annular artifacts are present in the initial medical image sequence, their radius can be determined using an annular artifact detection algorithm.

[0082] For example, the reconstruction algorithm can be the FDK (Feldkamp-Davis-Kress) reconstruction algorithm. The FDK reconstruction algorithm converts two-dimensional projection data into three-dimensional volume data, performs a Fourier transform, and filters the three-dimensional volume data to remove noise. Finally, an inverse Fourier transform is performed on the filtered data to obtain the reconstructed medical image sequence.

[0083] S120. Based on the radius of the annular artifact, determine the artifact slice image sequence among the multiple slice images included in the initial medical image sequence.

[0084] S130. Remove the ring artifacts from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

[0085] The artifact slice image sequence includes at least one target slice image containing annular artifacts. The artifact slice image sequence can be a collection of slice images containing annular artifacts. The radius of the annular artifact is related to the geometry and imaging parameters of the medical device detector. For example, the radius of annular artifacts in cone-beam CT depends on factors such as detector width, elliptic aspect ratio, and rotation center offset.

[0086] In some cases, medical image sequences can be used for diagnosis and treatment planning, such as detecting tumors, assessing neurological disorders, and examining cardiovascular diseases. However, ring artifacts can interfere with doctors' diagnosis of patients, making their elimination necessary. After determining the radius of the ring artifact, the slice number containing the artifact can be identified based on the geometric relationship between the size of different fields of view of the detector and the radius of the artifact on the reconstructed slice. By identifying the image sequences containing the artifact, artifact elimination can be more targeted and accurate.

[0087] Specifically, based on the radius of the annular artifact, target slice images containing annular artifacts can be identified from the initial medical image sequence. Multiple target slice images containing artifacts can constitute an artifact slice image sequence. Then, according to a correction algorithm, the annular artifacts can be eliminated from the target slice images included in the artifact slice image sequence to obtain the target medical image sequence. For example, the presence of circles can be detected directly by applying a Hough transform layer by layer to each slice image to determine whether annular artifacts exist and to determine the radius of the annular artifacts. Alternatively, each slice image can be first converted to polar coordinates, and then a Hough transform can be applied layer by layer to each slice image to detect whether straight lines exist and to determine whether annular artifacts exist and to determine the radius of the annular artifacts.

[0088] In the above embodiments, when annular artifacts exist in the initial medical image sequence, the radius of the annular artifacts in the initial medical image sequence is obtained. Based on the radius of the annular artifacts, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence. The annular artifacts are then eliminated from the artifact slice image sequence to obtain a target medical image sequence free of annular artifacts. This embodiment does not employ a long-term static detector placement method but proposes a novel image post-processing approach. By automatically identifying the artifact slice image sequence from the multiple slice images included in the initial medical image sequence and eliminating the annular artifacts from the artifact slice image sequence, the elimination of annular artifacts in slice images is automated, improving image quality and realism. Furthermore, by automatically identifying the artifact slice image sequence based on the radius of the annular artifacts, the removal efficiency of annular artifacts can be improved.

[0089] In some implementations, the initial medical image sequence corresponds to reconstructed imaging parameters. These reconstructed imaging parameters represent the parameters used to reconstruct the initial medical image sequence. Determining the artifact slice image sequence among multiple slice images included in the initial medical image sequence based on the radius of the annular artifact can include: determining the artifact slice image sequence among multiple slice images included in the initial medical image sequence based on the radius of the annular artifact and the reconstructed imaging parameters.

[0090] In medical imaging, reconstructed imaging parameters refer to a set of parameters used to generate the initial medical image sequence. These parameters may include the total number of reconstructed slices, pixel spacing, slice thickness, etc.

[0091] In some cases, reconstructed imaging parameters can directly affect the quality and accuracy of the initial medical image sequence. In medical imaging, different imaging techniques and application scenarios require different reconstructed imaging parameter settings. Specifically, when annular artifacts exist in the initial medical image sequence, to eliminate them, it is first necessary to obtain the parameters used to reconstruct the initial medical image sequence—the reconstructed imaging parameters—and determine the radius of the annular artifact. Next, based on the radius of the annular artifact and the reconstructed imaging parameters, multiple target slice images containing the artifact can be identified from the multiple slice images included in the initial medical image sequence. Finally, the artifact slice image sequence is constructed using these multiple target slice images.

[0092] In the above embodiments, based on the radius of the ring artifact and the reconstructed imaging parameters, the artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence. By determining the artifact slice image sequence containing the ring artifact, a data basis can be provided for subsequent elimination of the ring artifact, which is beneficial to improving the quality of medical images.

[0093] In some implementations, please refer to Figure 2a The reconstructed imaging parameters include the total number of reconstructed slices, pixel spacing, and slice thickness. Determining the artifact slice image sequence from the multiple slice images included in the initial medical image sequence, based on the radius of the annular artifact and the reconstructed imaging parameters, may include the following steps:

[0094] S210. Determine the imaging field size that causes the ring artifact based on the radius of the ring artifact.

[0095] The imaging field of view size can be the size of the detector used to detect the area to be detected. For example, please refer to [link to example]. Figure 2b The detector size can be 20×20, but since the area to be detected is small, the required detector size can be the area enclosed by rectangle 210, which can be 16×16. The imaging field of view size can also be the area enclosed by rectangle 210. The imaging field of view size is closely related to the area being detected. Specifically, the imaging field of view size should match the size of the area being detected to ensure complete coverage of the area to be detected during scanning and to obtain sufficient resolution and image quality.

[0096] In some cases, when the area to be examined is small, reducing the imaging field of view can enlarge the area for display on the medical device. By reducing the imaging field of view, more detailed information about the area can be obtained, allowing for more precise location of the lesion. This enables doctors to make more accurate diagnoses during surgery and helps them better understand the surrounding structures, improving surgical precision and safety.

[0097] It should be noted that when a detector is in operation for a period of time and its imaging field of view is smaller than the detector itself, if detection continues through that imaging field of view, the hardware performance of the detector's imaging field of view will be affected, and the performance of that area will also change. Therefore, the response between the detector's imaging field of view and other areas outside the imaging field of view will be inconsistent. This inconsistency in response will produce ring artifacts. The radius of the ring artifact is related to the detector's imaging field of view. Generally, when the detector's imaging field of view is small, the detector's inhomogeneity is small, resulting in a smaller artifact radius. Conversely, when the detector's imaging field of view is large, the detector's inhomogeneity is large, resulting in a larger artifact radius.

[0098] Specifically, the detector's imaging field of view (Field of View) is used to acquire medical images of the area to be examined. The Field of View determines the size of the ring artifact. Therefore, to eliminate the ring artifact, its radius is first determined. Then, based on the geometric relationship between the ring artifact radius and the Field of View, the Field of View size that causes the ring artifact is determined.

[0099] In some implementations, the imaging field of view (Field of View) corresponds to different radii of different annular artifacts. Therefore, a relationship between the radius of the annular artifact and the Field of View can be established. After obtaining the radius of the annular artifact, it can be used to search within the relationship data to determine the Field of View that caused the annular artifact.

[0100] S220. Determine the starting slice number and ending slice number of the ring artifact based on the imaging field size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness.

[0101] S230. Take the slice images located between the starting slice number and the ending slice number as the artifact slice image sequence.

[0102] The total number of reconstructed slices can be the total number of slice images included in the initial medical image sequence. Pixel spacing can be the distance between adjacent pixels. Slice thickness can be the thickness of each slice image in the initial medical image sequence. Smaller slice thickness results in higher image resolution, allowing for clearer display of details and structures, but it also increases scan time and radiation dose. Therefore, the choice of slice thickness requires a trade-off based on specific circumstances; different scan purposes and locations will necessitate different slice thicknesses. For example, for head scans, slice thickness is generally smaller, typically between 0.5-1.0 mm, to better display brain structures; while for lung scans, slice thickness can be appropriately increased, generally between 1.0-2.0 mm, to reduce the number of images and radiation dose. In summary, the slice thickness for medical images needs to be selected based on factors such as the scan purpose, location, and required resolution to obtain optimal image quality and diagnostic results.

[0103] Specifically, to eliminate ring artifacts, it is necessary to determine the slice image number where the ring artifact is located. By determining the slice image where the ring artifact is located, the ring artifact can be eliminated more accurately. Therefore, the starting slice number and ending slice number where the ring artifact exists can be determined based on the geometric relationship between the imaging field of view size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness. Slice images located between the starting slice number and the ending slice number can all be considered as slice images containing artifacts. Therefore, slice images located between the starting slice number and the ending slice number can be considered as the artifact slice image sequence.

[0104] In the above embodiments, the imaging field of view size is determined based on the radius of the ring artifact. The starting and ending slice numbers where the ring artifact exists are determined based on the imaging field of view size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness. Slice images located between the starting and ending slice numbers are taken as the artifact slice image sequence. By determining the starting and ending slice numbers where the ring artifact is located, the interval where the ring artifact is located can be determined, providing a data basis for subsequent ring artifact removal.

[0105] In some implementations, the initial medical image sequence corresponds to an active location parameter, which indicates the location of the radiation source. Determining the artifact slice image sequence from multiple slice images included in the initial medical image sequence based on the radius of the annular artifact and the reconstructed imaging parameters may include: determining the artifact slice image sequence from multiple slice images included in the initial medical image sequence based on the radius of the annular artifact, the reconstructed imaging parameters, and the active location parameter.

[0106] The source location parameter indicates the location of the radiation source. The location of the radiation source refers to the origin or source of radioactive energy, such as radioactive material or electromagnetic radiation. A radiation source is a radioactive material or other substance capable of producing radiation, such as an X-ray generator. Radiation sources release electromagnetic waves or particle beams with powerful penetrating and lethal force, posing a potential risk to human health.

[0107] Specifically, to eliminate ring artifacts, it is necessary to determine the artifact slice image sequence. By determining the artifact slice image sequence, the ring artifact can be eliminated more accurately. Therefore, the artifact slice image sequence can be determined from the multiple slice images included in the initial medical image sequence based on the geometric relationship between the radius of the ring artifact, the reconstructed imaging parameters, and the source location parameters.

[0108] In the above embodiments, based on the radius of the ring artifact, reconstructed imaging parameters, and source location parameters, a sequence of artifact slice images is determined from multiple slice images included in the initial medical image sequence. This provides a data foundation for subsequent ring artifact elimination, improving the quality of medical images and the accuracy of doctors' diagnoses.

[0109] In some implementations, the starting slice number and the ending slice number are determined using the following formula:

[0110]

[0111]

[0112] Wherein, SlicesFrames is the total number of slice images in the initial medical image sequence, SOD is the vertical distance between the X-ray source and the rotation center of the imaging device, SID is the vertical distance between the X-ray source and the center of the detector, F is the imaging field of view size, pixelpitch is the pixel pitch, SLicesThickness is the slice thickness, zs is the starting slice number, and ze is the ending slice number.

[0113] In this system, the source location is the focal point of the X-ray tube head, where the X-rays are generated. The rotation center can be the geometric center point between the source location and the detector's rotation axis, or it can be a fixed point within the medical device. The rotation center can also be the axis around which the object or patient being examined rotates during the imaging process. During imaging, the object or patient rotates around the rotation center, and the X-ray beam emitted from the source location passes through it and is finally received by the detector. Through rotational imaging, the medical device can acquire three-dimensional structural information of the object or patient being examined.

[0114] Specifically, please refer to Figure 3S is the source location, Z is the location of the rotation center, rectangle 310 can be the detector, rectangle 320 can be the imaging field of view size, F can be the size of the imaging field of view size, and M can be the total number of pixels in the slice image containing ring artifacts. The rotation center of the medical device can be the center position of the intermediate slice image in the initial image sequence, and the slice images are arranged sequentially based on the rotation center. According to the principle of triangle similarity, it can be concluded that... From this, we can deduce... Based on the thickness of each slice image, the formula can be used. Calculate the number of slice images containing artifacts. According to the formula... The number of slice images with ring artifacts to the right or left of the central slice image can be calculated. According to the formula... The number of slice images to the right or left of the central slice image can be calculated. Then, the relationship between the number of slice images on each side of the central slice image and the number of slice images with artifacts on each side of the central slice image can be established. The starting slice number zs where the annular artifact exists can be determined. This is based on the quantitative relationship between the number of slice images on both sides of the central slice image and the number of slice images on both sides of the central slice image that contain artifacts. The termination slice number ze, where the ring artifact is identified, can be determined.

[0115] For example, the initial image sequence can be 500 images, so the sequence number of the intermediate slice image can be determined to be 250, according to the formula. It can be determined that there are 150 slices with ring artifacts on both sides of the middle slice image. Therefore, subtracting 150 from the sequence number of the middle slice image determines that the starting slice sequence number with ring artifacts is 100. Adding 150 to the sequence number of the middle slice image determines that the ending slice sequence number with ring artifacts is 400.

[0116] In some implementations, the sliced ​​image is located in a Cartesian coordinate system. See also... Figure 4a The method for determining the radius of a ring artifact may include the following steps:

[0117] S410. Convert at least some slice images in the initial medical image sequence to polar coordinates to obtain a polar coordinate image sequence.

[0118] S420. Determine the radius of the ring artifact based on the image sequence in polar coordinates.

[0119] In polar coordinates, each point is defined by a polar radius and a polar angle. The polar radius represents the distance from the point to the origin, and the polar angle represents the angle between the point and the polar axis. The polar axis is a defined straight line. The polar angle can be positive or negative and is usually expressed in degrees or radians.

[0120] In some cases, transforming the artifact enhancement image from Cartesian coordinates to polar coordinates can reduce the difficulty of artifact extraction and processing, better display the symmetry of annular artifacts, and more intuitively represent the spatial distribution of the artifact enhancement image in polar coordinates. In this embodiment, the horizontal axis in polar coordinates can represent the radius of the annular artifact, and the vertical axis can represent the angle of the annular artifact.

[0121] Specifically, not every slice image in the initial medical image sequence contains annular artifacts. Therefore, a subset of slice images can be selected from the initial medical image sequence. The center of each selected slice image can be set as the origin of a Cartesian coordinate system. The position of any pixel in the slice image in the Cartesian coordinate system is set as (x, y), and its corresponding polar coordinate point is set as (r, θ). The horizontal axis in the polar coordinate system represents the radius of the annular artifact, and the vertical axis represents the angle of the annular artifact, according to the following formula:

[0122] r = sqrt(x 2 +y 2 )

[0123]

[0124] The pixels (x, y) of a sliced ​​image in Cartesian coordinates can be converted to polar coordinates (r, θ). By transforming each pixel in the sliced ​​image from Cartesian to polar coordinates, a polar coordinate image sequence can be obtained. Image processing operations can then be performed on this polar coordinate image sequence to determine the radius of the annular artifact.

[0125] For example, please refer to Figure 4b , Figure 4b This is a single slice image in Cartesian coordinates. See also... Figure 4c , Figure 4c A single slice image in Cartesian coordinate system Figure 4b The image is transformed to polar coordinates.

[0126] In the above embodiments, at least some slice images in the initial medical image sequence are converted to polar coordinates to obtain a polar coordinate image sequence. The radius of the ring artifact is determined based on the polar coordinate image sequence. By determining the radius of the ring artifact, the slice image number where the ring artifact is located can be determined. By determining the slice image number of the ring artifact, the ring artifact can be eliminated.

[0127] In some implementations, please refer to Figure 4d Obtaining the radius of the annular artifact in the initial medical image sequence can include the following steps:

[0128] S402. Convert at least some slice images in the initial medical image sequence to polar coordinates to obtain a polar coordinate image sequence.

[0129] Specifically, not every slice image in the initial medical image sequence contains annular artifacts; therefore, a subset of slice images can be selected from the initial medical image sequence. The center of each selected slice image can be set as the origin of a Cartesian coordinate system, according to the following formula:

[0130] r = sqrt(x 2 +y 2 )

[0131]

[0132] The pixels (x, y) of a sliced ​​image in Cartesian coordinates can be converted to polar coordinates (r, θ). By transforming each pixel in the sliced ​​image from Cartesian coordinates to polar coordinates, a sequence of polar coordinate images can be obtained.

[0133] S404. Based on the image sequence in polar coordinate system, obtain the first gray-level distribution curve.

[0134] S406. Extract the gray-level distortion points from the first gray-level distribution curve, and use the radius corresponding to the gray-level distortion points as the radius of the ring artifact.

[0135] The first grayscale distribution curve represents the average pixel value of at least a portion of the sliced ​​image at different radii in the polar coordinate system. Grayscale distortion points refer to situations in digital image processing where the grayscale values ​​of certain pixels in an image are abnormal or distorted.

[0136] In some cases, ring artifacts can cause abnormal or distorted grayscale distribution, resulting in noticeable grayscale distortion points in the image. The grayscale value of these distortion points may differ from that of the surrounding normal area, and may be significantly higher.

[0137] Specifically, pixel processing is performed on each polar coordinate image in the polar coordinate image sequence to obtain a first gray-level distribution curve. Since circular artifacts exist in the polar coordinate image sequence, gray-level distortion points exist in the first gray-level distribution curve. These gray-level distortion points are extracted from the first gray-level distribution curve, and the radius corresponding to each distortion point is used as the radius of the circular artifact.

[0138] In the above embodiments, at least some slice images in the initial medical image sequence are converted to polar coordinates to obtain a polar coordinate image sequence. Based on the polar coordinate image sequence, a first gray-level distribution curve is obtained, and gray-level distortion points in the first gray-level distribution curve are extracted. The radius corresponding to the gray-level distortion points is used as the radius of the ring artifact. By determining the radius of the ring artifact, the slice image number where the ring artifact is located can be determined. By determining the slice image number of the ring artifact, the ring artifact can be eliminated.

[0139] In some implementations, please refer to Figure 5a Determining the radius of a ring artifact based on a polar coordinate system image sequence may include the following steps:

[0140] S510. Based on the number of pixels in the polar coordinates, perform pixel averaging on any polar coordinate image in the polar coordinate image sequence to obtain the mean pixel data corresponding to the polar coordinate image sequence.

[0141] Specifically, in a polar coordinate system image sequence, the mean pixel data corresponding to that polar coordinate system image sequence can be obtained by averaging all pixels in each column along the polar angular coordinate direction. The horizontal axis of the image displaying the mean pixel data is the radius, and the vertical axis is the number of slices.

[0142] For example, please refer to Figure 5c A polar coordinate system image sequence can contain polar coordinate system images P0, P1, and P2. Polar coordinate system image P0 contains pixels P... 00 P 01 P 02 P 03 P 04 P 05 The polar coordinate system image P1 contains pixels P 10 P 11 P 12 P 13 P 14 P 15 The polar coordinate system image P2 contains pixels P 20 P 21 P 22 P 23 P 24 P 25 The polar coordinate system image P0 contains the pixels P. 00 With P0 containing pixel P 03 Adding the two numbers together yields the pixel and D values. 01 , to pixel and D 01 Dividing by 2 yields the mean pixel data A0, and the polar coordinate image P0 contains the pixel P... 01 With P0 containing pixel P04 Adding the two numbers together yields the pixel and D values. 02 , to pixel and D 02 Dividing by 2 yields the mean pixel data B0, and the polar coordinate image P0 contains the pixels P 02 With P0 containing pixel P 05 Adding the two numbers together yields the pixel and D values. 03 , to pixel and D 03 Dividing by 2 yields the mean pixel data C0. The pixel values ​​P contained in the polar coordinate image P1 are then... 10 With P1 containing pixel P 13 Adding the two numbers together yields the pixel and D values. 11 , to pixel and D 11 Dividing by 2 yields the mean pixel data A1, and the pixel P contained in the polar coordinate image P1 is... 11 With P1 containing pixel P 14 Adding the two numbers together yields the pixel and D values. 12 , to pixel and D 12 Dividing by 2 yields the mean pixel data B1, and the polar coordinate image P1 contains the pixel P... 12 With P1 containing pixel P 15 Adding the two numbers together yields the pixel and D values. 13 , to pixel and D 13 Dividing by 2 yields the mean pixel data C1. The pixel values ​​P contained in the polar coordinate image P2 are then used to calculate the mean pixel data. 20 With P2 containing pixel P 23 Adding the two numbers together yields the pixel and D values. 21 , to pixel and D 21 Dividing by 2 yields the mean pixel data A2, and the polar coordinate image P2 contains the pixel P... 21 With P2 containing pixel P 24 Adding the two numbers together yields the pixel and D values. 22 , to pixel and D 22 Dividing by 2 yields the mean pixel data B2, which in turn represents the pixel P contained in the polar coordinate image P2. 22 With P2 containing pixel P 25 Adding the two numbers together yields the pixel and D values. 23 , to pixel and D 23 Dividing by 2 yields the mean pixel data C2. See also... Figure 5d By averaging the pixels of polar coordinate images P0, P1, and P2, the mean pixel data corresponding to the polar coordinate image sequence can be obtained. Figure 5d The horizontal axis represents the radius, and the vertical axis represents the number of slices.

[0143] S520. Based on the number of polar coordinate system images, the mean pixel data at any polar radius coordinate is averaged to obtain the first gray-level distribution curve.

[0144] S530. Determine the radius of the annular artifact based on the first grayscale distribution curve.

[0145] In some cases, grayscale refers to the brightness value of each pixel in an image, typically expressed as an integer value between 0 and 255, where 0 represents black and 255 represents white. The higher the grayscale value, the brighter the pixel. Artifacts are more likely to appear when grayscale values ​​transition abruptly.

[0146] Specifically, after obtaining the mean pixel data corresponding to the polar coordinate image sequence, the non-zero mean pixel data can be averaged column by column along the slice number direction to obtain the first gray-level distribution curve along the radius direction. Then, the radius of the annular artifact can be determined based on the first gray-level distribution curve.

[0147] For example, please refer to Figure 5d When the mean pixel data A0, A1, and A2 are all non-zero, the mean pixel data A0, A1, and A2 can be added together, and then the sum of the mean pixel data A0, A1, and A2 can be divided by 3. When the mean pixel data B1 in the mean pixel data B0, B1, and B2 is zero, the mean pixel data B0 and B2 can be added together, and then the sum of the mean pixel data B0, B1, and B2 can be divided by 2. When the mean pixel data C0, C1, and C2 are all non-zero, the mean pixel data C0, C1, and C2 can be added together, and then the sum of the mean pixel data C0, C1, and C2 can be divided by 3. This allows pixel averaging of any polar coordinate system image in a polar coordinate system image sequence. Please refer to [link / reference]. Figure 5e The first grayscale distribution curve can be obtained by performing pixel averaging on the polar coordinate system image.

[0148] In the above embodiments, the pixel averaging process is performed on any polar coordinate system image in the polar coordinate system image sequence based on the number of pixels at the polar angle coordinates to obtain the mean pixel data corresponding to the polar coordinate system image sequence. The mean pixel data at any polar radius coordinate is averaged based on the number of polar coordinate system images to obtain the first gray-level distribution curve. The radius of the ring artifact can be determined according to the first gray-level distribution curve.

[0149] In some implementations, please refer to Figure 5f Obtaining the first grayscale distribution curve based on a polar coordinate system image sequence may include the following steps:

[0150] S502. Based on the number of pixels in polar coordinates, perform pixel averaging on any polar coordinate image in the polar coordinate image sequence in the angular direction to obtain the mean pixel data corresponding to different radii of each slice image in the polar coordinate image sequence.

[0151] S504. Based on the mean pixel data corresponding to different radii of each slice image, average the data along the slice direction to obtain the first grayscale distribution curve.

[0152] Specifically, in a polar coordinate system image sequence, averaging all pixels in each column along the angular direction yields the mean pixel data corresponding to different radii for each slice image in the sequence. The horizontal axis of the image displaying the mean pixel data represents the radius, and the vertical axis represents the number of slices. After obtaining the mean pixel data corresponding to different radii for each slice image in the polar coordinate system image sequence, the non-zero mean pixel data can be averaged column-by-column along the slice number direction to obtain the first grayscale distribution curve. For an example, please refer to [link to example]. Figure 5b , Figure 5b This is an image obtained by averaging the image pixel by pixel in the angular direction in polar coordinates.

[0153] In the above embodiments, pixel averaging is performed on any polar coordinate image in the polar coordinate image sequence along the angular direction based on the number of pixels in polar angular coordinates. This yields the mean pixel data corresponding to different radii for each slice image in the polar coordinate image sequence. Based on the mean pixel data corresponding to different radii for each slice image, averaging is performed along the slice direction to obtain a first grayscale distribution curve. The grayscale distortion point can be determined based on the first grayscale distribution curve. Furthermore, the radius of the ring artifact can be determined based on the location of the grayscale distortion point.

[0154] In some implementations, please refer to Figure 6 Extracting gray-level distortion points from the first gray-level distribution curve may include the following steps:

[0155] S610. Smooth the first grayscale distribution curve to obtain the second grayscale distribution curve.

[0156] Smoothing can be achieved through filtering algorithms, such as moving average filtering, median filtering, and Gaussian filtering.

[0157] In some cases, when a pixel transitions from black to white, artifacts can appear in the transition region if the transition is not smooth. Therefore, in digital image processing, smoothing methods are needed to reduce artifacts caused by gray-level transitions. Smoothing the first gray-level distribution curve can smooth out the areas with large fluctuations, or what can be understood as spiky parts. Then, image processing based on the smoothed gray-level distribution curve and the first gray-level curve can highlight the locations of ring-shaped artifacts.

[0158] Specifically, the first gray-level distribution curve is smoothed by a filtering function. By smoothing the first gray-level distribution curve, the prominent parts of the first gray-level distribution curve can be made smoother, thus obtaining the second gray-level distribution curve.

[0159] For example, the L0 norm filtering method can be used to smooth the first gray-level distribution curve to obtain the second gray-level distribution curve.

[0160] S620. Subtract the first gray level distribution curve from the second gray level distribution curve to obtain the target gray level distribution curve.

[0161] S630. Take the peak of the target grayscale distribution curve as the grayscale distortion point.

[0162] In some cases, the gray value of the location of the ring artifact will be greater than the gray value of the surrounding normal tissue. Therefore, the location where the data difference between the second gray-scale distribution curve and the first gray-scale distribution curve is the largest can be considered as the location of the artifact.

[0163] Specifically, the target grayscale distribution curve can be obtained by subtracting the second grayscale distribution curve from the first grayscale distribution curve. The peak value of this target grayscale distribution curve can then be determined. The location of the peak value can be the position where the grayscale value changes the most, and this position can be considered the grayscale distortion point.

[0164] In the above embodiments, the first gray-level distribution curve is smoothed to obtain a second gray-level distribution curve. A subtraction operation is performed between the first and second gray-level distribution curves to obtain a target gray-level distribution curve. The peak value of the target gray-level distribution curve is taken as the gray-level distortion point. By determining the location of the gray-level distortion point, the radius of the ring artifact can be determined. By determining the radius of the ring artifact, the slice image number where the ring artifact is located can be determined. By determining the slice image number of the ring artifact, the ring artifact can be eliminated.

[0165] In some implementations, please refer to Figure 7aRemoving annular artifacts from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence may include the following steps:

[0166] S710. Perform pixel-by-pixel averaging on the target slice image included in the artifact slice image sequence to obtain the initial artifact enhancement image.

[0167] The initial artifact-enhanced image is located in a Cartesian coordinate system. The Cartesian coordinate system is a commonly used coordinate system in mathematics, also known as a rectangular coordinate system. It consists of a number line and a right angle, used to describe the position of points in a plane or space. In a two-dimensional Cartesian coordinate system, the position of a point in a plane is determined by two mutually perpendicular number lines (called the x-axis and y-axis). In a Cartesian coordinate system, the position of each point can be represented by an ordered pair.

[0168] Specifically, the artifact slice image sequence contains multiple target slice images. By averaging each target slice image in the artifact slice image sequence pixel by pixel, an initial artifact enhancement image can be obtained.

[0169] For example, please refer to Figure 7b The artifact slice image sequence can contain three target slice images P4, P5, and P6. Target slice image P4 contains pixels P... 40 P 41 P 42 The target slice image P5 contains pixels P 50 P 51 P 52 The target slice image P6 contains pixels P 60 P 61 P 62 The target slice image P4 contains pixels P 40 The target slice image P5 contains pixels P 50 The target slice image P6 contains pixels P 60 Adding the pixels together yields a sum of pixels (P0 and W0). Dividing this sum by 3 performs pixel averaging. The target slice image P4 contains pixels P0. 41 The target slice image P5 contains pixels P 51 The target slice image P6 contains pixels P 61 Adding the pixels together yields a sum of pixels and W1. Dividing this sum by 3 performs pixel averaging. The target slice image P4 contains pixels P... 42 The target slice image P5 contains pixels P 52 The target slice image P6 contains pixels P 62Adding the pixels together yields the sum of the pixels and W2. Dividing the sum of the pixels and W2 by 3 performs pixel averaging. By averaging each pixel of the target slice images P4, P5, and P6, the initial artifact enhancement image corresponding to the artifact slice image sequence can be obtained.

[0170] S720. Transform the initial artifact enhancement image to polar coordinates to obtain the initial linear artifact image.

[0171] In some cases, transforming the initial artifact enhancement image from Cartesian coordinates to polar coordinates can reduce the difficulty of artifact extraction and processing, better display the symmetry of annular artifacts, and more intuitively represent the spatial distribution of the initial artifact enhancement image in polar coordinates. Annular artifacts in Cartesian coordinates transform into striped artifacts with simpler and more distinct geometric features in the vertical direction, without a significant change in image resolution.

[0172] Specifically, the center point of the initial artifact enhancement image can be set as the origin of the Cartesian coordinate system. The position of any pixel in the initial artifact enhancement image in the Cartesian coordinate system is set as (x, y), and its corresponding polar coordinate point is set as (r, θ). In the polar coordinate system, the horizontal axis r represents the radius of the annular artifact, and the vertical axis θ represents the angle of the annular artifact, according to the following formula:

[0173]

[0174]

[0175] The pixel location (x, y) can be converted to polar coordinates (r, θ). By transforming each pixel of the initial artifact enhancement image from Cartesian coordinates to polar coordinates, an initial linear artifact image in the vertical direction can be obtained.

[0176] For example, the initial artifact enhancement image can be converted from Cartesian coordinates to polar coordinates using bilinear interpolation to obtain the initial linear artifact image.

[0177] S730. Filter the initial linear artifact image to obtain the target linear artifact image.

[0178] Among them, filtering can refer to the processing of a signal by applying a filter to change the signal’s frequency response or amplitude characteristics, thereby removing noise, smoothing the signal, enhancing certain frequency components, and making the signal clearer and more accurate.

[0179] Specifically, in the angular direction, the initial linear artifact image is filtered column by column. The filtering process can eliminate noise in the initial linear artifact image to obtain the intermediate linear artifact image. Then, in the radial direction, the intermediate linear artifact image can be filtered row by row to obtain the target linear artifact image.

[0180] For example, the filtering process and the one-dimensional filtering process can be median filtering or mean filtering. In the radial direction, the kernel function of the one-dimensional filter can be set according to the size of the radius. Different radius sizes correspond to different kernel functions. As the radius value changes in the radial direction, the kernel function of the one-dimensional filter will also change accordingly. For each radius value, there is a corresponding one-dimensional filter to perform the filtering process.

[0181] S740. Transform the target linear artifact image to the Cartesian coordinate system to obtain the target artifact enhanced image.

[0182] S750. Based on the target artifact enhancement image, perform artifact removal on the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

[0183] Specifically, the position of any pixel in the target linear artifact image in polar coordinates is set to (r, θ), and its corresponding Cartesian coordinates are set to (x, y), according to the following formula:

[0184] x = rcosθ

[0185] y = rsinθ

[0186] The polar coordinates (r, θ) of a pixel can be converted to Cartesian coordinates (x, y). By transforming each pixel of the pseudo-target linear artifact image from polar coordinates to Cartesian coordinates, an enhanced image of the target artifact can be obtained. Then, image processing operations can be performed on the target slice images included in the artifact slice image sequence based on the enhanced image of the target artifact to eliminate the ring artifacts and obtain the target medical image sequence.

[0187] For example, a bilinear interpolation algorithm can be used to transform the target linear artifact image from polar coordinates to Cartesian coordinates to obtain an enhanced target artifact image.

[0188] In the above embodiments, pixel-by-pixel averaging is performed on the target slice images included in the artifact slice image sequence to obtain an initial artifact enhancement image. The initial artifact enhancement image is then transformed to polar coordinates to obtain an initial linear artifact image. Filtering is performed on the linear artifact image to obtain a target linear artifact image. The target linear artifact image is then transformed to Cartesian coordinates to obtain a target artifact enhancement image. Artifact removal is then performed on the target slice images included in the artifact slice image sequence based on the target artifact enhancement image to obtain a target medical image sequence. By eliminating the target artifact enhancement image from the artifact slice image sequence, the effect of eliminating ring artifacts can be achieved, improving the quality of medical images, enhancing the accuracy of doctors' diagnoses, and reducing the probability of misdiagnosis.

[0189] In some implementations, artifact removal is performed on target slice images included in the artifact slice image sequence based on the target artifact enhancement image to obtain a target medical image sequence. This includes: subtracting the corresponding pixels of any target slice image in the artifact slice image sequence from the target artifact enhancement image to obtain an artifact-removed slice image corresponding to any target slice image; and replacing any target slice image with the artifact-removed slice image to obtain the target medical image sequence. Alternatively, artifact removal pixel data is determined based on artifact adjustment data and the target artifact enhancement image; corresponding pixels are subtracted from the artifact removal pixel data based on the artifact slice image sequence to obtain an artifact-removed slice image corresponding to any target slice image; and replacing any target slice image with the artifact-removed slice image to obtain the target medical image sequence.

[0190] Among them, the artifact-removed slice image can be a target slice image with the ring artifact removed. Replacing the target slice image with the ring artifact with the corresponding artifact-removed slice image can yield the target medical image sequence.

[0191] Specifically, the artifact slice image sequence contains multiple target slice images, each with a corresponding target artifact enhancement image. By subtracting the corresponding pixel value from any target slice image and its corresponding target artifact enhancement image, the artifact can be removed from the target slice image, resulting in an artifact-free slice image for any target slice image. Replacing any target slice image with the artifact-free slice image yields the target medical image sequence. This implementation achieves the correction of ring artifacts.

[0192] The artifact adjustment data can be determined based on the actual situation. For example, if the intensity of the ring artifacts on the target slice image M is significant, the artifact adjustment data for target slice image M can be set to 1. This allows multiplying each pixel of the target artifact enhancement image corresponding to target slice image M by 1 to obtain the artifact removal pixel data. In other words, the corresponding pixel subtraction operation can be performed between the target slice image M and its corresponding artifact removal pixel data to obtain the artifact-removed slice image. Similarly, if the intensity of the ring artifacts on target slice image N is weaker than that on target slice image M, the adjustment data can be set according to the degree of difference between the ring artifacts on target slice image N and target slice image M. For example, the artifact adjustment data for target slice image N can be set to 1 / 3. This allows multiplying each pixel of the target artifact enhancement image corresponding to target slice image N by 1 / 3 to obtain the artifact removal pixel data. In other words, the corresponding pixel subtraction operation can be performed between the target slice image N and its corresponding artifact removal pixel data to obtain the artifact removal slice image.

[0193] The intensity of ring artifacts displayed on the target slice images in a sequence of artifact slice images is inconsistent. Therefore, artifact adjustment data for each target slice image can be set according to the intensity of the ring artifact on each target slice image. Then, the artifact removal pixel data for each target slice image can be determined by multiplying the artifact adjustment data and the corresponding artifact enhancement image. Subtracting the corresponding pixel data from the artifact removal pixel data for each target slice image in the artifact slice image sequence removes the artifact from the target slice image, resulting in an artifact-removed slice image for any target slice image. Replacing any target slice image with the artifact-removed slice image yields the target medical image sequence. This embodiment achieves the correction of ring artifacts.

[0194] In the above embodiments, a pixel-wise subtraction operation is performed between any target slice image in the artifact slice image sequence and the target artifact enhancement image to obtain an artifact-removed slice image corresponding to any target slice image; the artifact-removed slice image is then used to replace any target slice image to obtain the target medical image sequence. Alternatively, artifact removal pixel data is determined based on artifact adjustment data and the target artifact enhancement image, and a pixel-wise subtraction operation is performed between the artifact slice image sequence and the artifact removal pixel data to obtain an artifact-removed slice image corresponding to any target slice image; the artifact-removed slice image is then used to replace any target slice image to obtain the target medical image sequence. By performing pixel subtraction operations, ring artifacts can be removed while retaining detailed information in the initial medical image sequence, ensuring the integrity and accuracy of the original medical image.

[0195] This specification also provides a medical image processing method. An initial medical image sequence corresponds to reconstructed imaging parameters, which represent the parameters used to reconstruct the initial medical image sequence. These reconstructed imaging parameters include the total number of reconstructed slices, pixel spacing, and slice thickness. For example, please refer to... Figure 8 The medical image processing method may include the following steps:

[0196] S802. Obtain the initial medical image sequence.

[0197] Among them, there are ring artifacts in the initial medical image sequence.

[0198] S804. Convert at least some slice images in the initial medical image sequence to polar coordinates to obtain a polar coordinate image sequence.

[0199] S806. Based on the number of pixels in the polar coordinate system, perform pixel averaging on any polar coordinate system image in the polar coordinate system image sequence to obtain the mean pixel data corresponding to the polar coordinate system image sequence.

[0200] S808. The mean pixel data at any polar radius coordinate is averaged based on the number of polar coordinate images to obtain the first grayscale distribution curve.

[0201] S810. Smooth the first grayscale distribution curve to obtain the second grayscale distribution curve.

[0202] S812. Subtract the first gray level distribution curve from the second gray level distribution curve to obtain the target gray level distribution curve.

[0203] S814. Use the polar radius coordinates corresponding to the peak of the target grayscale distribution curve as the radius of the annular artifact.

[0204] S816. Determine the imaging field size that causes the ring artifact based on the radius of the ring artifact.

[0205] S818. Determine the starting slice number and ending slice number of the ring artifact based on the imaging field size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness.

[0206] Specifically, the starting slice number and the ending slice number are determined using the following formula:

[0207]

[0208]

[0209] Where zs is the starting slice number, ze is the ending slice number, SlicesFrames is the total number of slice images in the initial medical image sequence, F is the imaging field of view size, pixelpitch is the pixel pitch, SLicesThickness is the slice thickness, SOD is the vertical distance between the X-ray source and the rotation center of the imaging device, and SID is the vertical distance between the X-ray source and the center of the detector.

[0210] S820. The slice image located between the starting slice number and the ending slice number is taken as the target slice image in the artifact slice image sequence.

[0211] The artifact slice image sequence includes at least one target slice image containing a ring artifact.

[0212] S822. Perform pixel-by-pixel averaging on the target slice image included in the artifact slice image sequence to obtain the initial artifact enhancement image.

[0213] The initial artifact enhancement image is located in Cartesian coordinates.

[0214] S824. Transform the initial artifact enhancement image to polar coordinates to obtain the initial linear artifact image.

[0215] S826. Filter the linear artifact image to obtain the target linear artifact image.

[0216] S828. Transform the target linear artifact image to the Cartesian coordinate system to obtain the target artifact enhanced image.

[0217] S830. Subtract the corresponding pixels of any target slice image from the target artifact enhancement image in the artifact slice image sequence to obtain the artifact-removed slice image corresponding to any target slice image; replace any target slice image with the artifact-removed slice image to obtain the target medical image sequence.

[0218] This specification provides a medical image processing device 900. Please refer to [link / reference]. Figure 9The medical image processing device 900 includes: an initial sequence acquisition module 910, an artifact sequence determination module 920, and a ring artifact elimination module 930.

[0219] The initial sequence acquisition module 910 is used to acquire the radius of the annular artifact in the initial medical image sequence; wherein, the initial medical image sequence includes multiple slice images;

[0220] The artifact sequence determination module 920 is used to determine an artifact slice image sequence among multiple slice images included in the initial medical image sequence based on the radius of the annular artifact; wherein the artifact slice image sequence includes at least one target slice image containing the annular artifact;

[0221] The annular artifact removal module 930 is used to remove the annular artifact from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

[0222] For a detailed description of the medical image processing device, please refer to the description of the medical image processing method above, which will not be repeated here.

[0223] In some embodiments, a medical imaging device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method steps described above.

[0224] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in any of the above embodiments.

[0225] One embodiment of this specification provides a computer program product including instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0226] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

Claims

1. A medical image processing method, characterized in that, The method includes: Obtaining the radius of a ring artifact in an initial medical image sequence, wherein the initial medical image sequence includes multiple slice images, comprising: converting at least some slice images in the initial medical image sequence to a polar coordinate system to obtain a polar coordinate system image sequence; obtaining a first gray-level distribution curve based on the polar coordinate system image sequence; wherein the first gray-level distribution curve represents the average pixel value of the at least some slice images corresponding to different radii in the polar coordinate system; extracting gray-level distortion points in the first gray-level distribution curve, and using the radius corresponding to the gray-level distortion points as the radius of the ring artifact; Based on the radius of the annular artifact, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence; wherein, the artifact slice image sequence includes at least one target slice image containing the annular artifact; The annular artifact is removed from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

2. The method according to claim 1, characterized in that, The initial medical image sequence corresponds to reconstructed imaging parameters and source location parameters. The reconstructed imaging parameters represent the parameters used to reconstruct the initial medical image sequence; the source location parameters represent the location of the radiation source; determining the artifact slice image sequence from the multiple slice images included in the initial medical image sequence based on the radius of the annular artifact includes: Based on the radius of the annular artifact, the reconstructed imaging parameters, and the source location parameters, an artifact slice image sequence is determined from the multiple slice images included in the initial medical image sequence.

3. The method according to claim 2, characterized in that, The reconstructed imaging parameters include the total number of reconstructed slices, pixel spacing, and slice thickness; determining the artifact slice image sequence from the multiple slice images included in the initial medical image sequence based on the radius of the annular artifact and the reconstructed imaging parameters includes: The imaging field of view that causes the annular artifact is determined based on the radius of the annular artifact; The starting slice number and ending slice number where the annular artifact exists are determined based on the imaging field size, the total number of reconstructed slices, the pixel spacing, and the slice layer thickness. The slice images located between the starting slice number and the ending slice number are used as the artifact slice image sequence; The starting slice number and the ending slice number are determined using the following formulas: ； ； in, The starting slice number. The terminator is the termination slice number. The total number of slice images in the initial medical image sequence. For the imaging field of view size, For pixel spacing, For slice layer thickness, The vertical distance between the X-ray source and the center of rotation of the imaging device is denoted as . The vertical distance between the radiation source and the center of the detector is denoted as .

4. The method according to claim 1, characterized in that, Based on the polar coordinate system image sequence, the first gray-level distribution curve is obtained, including: Based on the number of pixels in polar coordinates, the average pixel data of each slice image in the polar coordinate system image sequence is obtained by averaging the pixels in the angular direction in any polar coordinate system image in the polar coordinate system image sequence at different radii. The first grayscale distribution curve is obtained by averaging the mean pixel data corresponding to different radii of each slice image along the slice direction.

5. The method according to claim 1, characterized in that, Extracting gray-level distortion points from the first gray-level distribution curve includes: The first grayscale distribution curve is smoothed to obtain the second grayscale distribution curve; The target grayscale distribution curve is obtained by subtracting the second grayscale distribution curve from the first grayscale distribution curve. The peak of the target grayscale distribution curve is taken as the grayscale distortion point.

6. The method according to claim 1, characterized in that, The step of removing the annular artifact from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence includes: The initial artifact enhancement image is obtained by performing pixel-by-pixel averaging on the target slice images included in the artifact slice image sequence; wherein, the initial artifact enhancement image is located in the Cartesian coordinate system; The initial artifact enhancement image is transformed to polar coordinates to obtain the initial linear artifact image; The target line artifact image is obtained by filtering the initial line artifact image. The target linear artifact image is transformed into a Cartesian coordinate system to obtain an enhanced target artifact image; Based on the target artifact enhancement image, artifact removal is performed on the target slice images included in the artifact slice image sequence to obtain the target medical image sequence.

7. The method according to claim 6, characterized in that, The process of removing artifacts from the target slice images included in the artifact slice image sequence based on the target artifact enhancement image to obtain a target medical image sequence includes: The target medical image sequence is obtained by subtracting the corresponding pixels of any target slice image in the artifact slice image sequence from the target artifact enhancement image; the artifact-removed slice image is then used to replace the target slice image to obtain the target medical image sequence; or... Based on the artifact adjustment data and the target artifact enhancement image, artifact removal pixel data is determined. A corresponding pixel subtraction operation is performed between the artifact slice image sequence and the artifact removal pixel data to obtain the artifact removal slice image corresponding to any target slice image. The target medical image sequence is obtained by replacing any target slice image with the artifact removal slice image.

8. A medical image processing device, characterized in that, The device includes: An initial sequence acquisition module is used to acquire the radius of annular artifacts in an initial medical image sequence. The initial medical image sequence includes multiple slice images, and the module includes: converting at least a portion of the slice images in the initial medical image sequence to a polar coordinate system to obtain a polar coordinate system image sequence; acquiring a first grayscale distribution curve based on the polar coordinate system image sequence; wherein the first grayscale distribution curve represents the average pixel value of the at least a portion of the slice images corresponding to different radii in the polar coordinate system; extracting grayscale distortion points from the first grayscale distribution curve, and using the radius corresponding to the grayscale distortion points as the radius of the annular artifact. The artifact sequence determination module is used to determine an artifact slice image sequence from multiple slice images included in the initial medical image sequence based on the radius of the annular artifact; wherein the artifact slice image sequence includes at least one target slice image containing the annular artifact; The ring artifact removal module is used to remove the ring artifact from the target slice image included in the artifact slice image sequence to obtain the target medical image sequence.

9. A medical imaging device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

Images