Method for segmenting intracranial blood vessels, electronic device, and storage medium

Abstract

The application discloses an intracranial blood vessel segmentation method, an electronic device and a storage medium. The intracranial blood vessel segmentation method comprises the following steps: acquiring a magnetic resonance blood vessel image to be segmented and a corresponding modality image; determining a brain tissue region image and a furrow region image according to the magnetic resonance blood vessel image and the modality image; performing blood vessel segmentation processing on the brain tissue region according to the brain tissue region image to obtain a brain tissue region blood vessel image; performing blood vessel segmentation processing on the furrow region according to the furrow region image to obtain a furrow region blood vessel image; and fusing the brain tissue region blood vessel image and the furrow region blood vessel image to obtain a target intracranial blood vessel image. The brain tissue region and the furrow region are subjected to blood vessel segmentation processing respectively, the obtained brain tissue region blood vessel image and the furrow region blood vessel image are fused, a more complete and accurate intracranial blood vessel segmentation result can be obtained, and the application range is wider without relying on sample data.

Description

Technical Field

[0001] This invention relates to the field of medical image processing, and in particular to a method for segmenting intracranial blood vessels, an electronic device, and a storage medium. Background Technology

[0002] Neurosurgery requires precise preoperative localization of the lesion, design and optimization of the surgical path, accurate intraoperative resection of the lesion area, and postoperative evaluation of the resection effect and recovery. Determining the location of intracranial blood vessels is crucial in neurosurgery. Effective intracranial vascular segmentation methods help to better distinguish intracranial blood vessels from surrounding tissues, allowing for clearer observation of intracranial blood vessels and thus providing more precise information for surgical planning.

[0003] Deep learning-based intracranial vessel segmentation is currently the mainstream method. It relies on a large amount of training data samples and a gold standard, continuously training neural networks to learn the characteristics of cerebral blood vessels and ultimately completing intracranial vessel segmentation. However, because creating the gold standard is extremely time-consuming, deep learning-based intracranial vessel segmentation methods are only suitable for scenarios with sufficient sample sizes and cannot be applied to scenarios with small sample sizes and significant differences between samples. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of existing deep learning-based intracranial vessel segmentation methods that cannot be applied to scenarios with small sample sizes and large differences between samples, and to provide an intracranial vessel segmentation method, electronic device and storage medium that do not rely on sample data and have a wider range of applications.

[0005] The present invention solves the above-mentioned technical problems through the following technical solution:

[0006] A first aspect of the present invention provides a method for segmenting intracranial blood vessels, comprising the following steps:

[0007] Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image;

[0008] Brain tissue region images and sulcus region images are determined based on the magnetic resonance angiography images and the modal images;

[0009] Based on the brain tissue region image, blood vessel segmentation processing of the brain tissue region is performed to obtain a blood vessel image of the brain tissue region;

[0010] Based on the image of the groove region, perform blood vessel segmentation processing on the groove region to obtain a blood vessel image of the groove region;

[0011] By fusing the vascular images of the brain tissue region and the vascular images of the groove region, a target intracranial vascular image is obtained.

[0012] Optionally, determining brain tissue region images based on the magnetic resonance angiography and the modal images specifically includes:

[0013] Remove non-brain tissue regions from the modal images;

[0014] The modal image excluding non-brain tissue regions is registered with the magnetic resonance angiography image to obtain a brain tissue region image.

[0015] Optionally, determining the groove region image based on the magnetic resonance vascular image and the modal image specifically includes:

[0016] Morphological processing was performed on the brain tissue region images to obtain intracranial images;

[0017] The sulcus region image is determined based on the brain tissue region image and the intracranial image.

[0018] Optionally, the step of performing blood vessel segmentation processing on the brain tissue region image to obtain a blood vessel image of the brain tissue region specifically includes:

[0019] The intracranial image and the magnetic resonance angiography image are masked to obtain a first image;

[0020] The first image is processed by blood vessel segmentation to obtain an initial intracranial blood vessel image;

[0021] The brain tissue region image is masked with the initial intracranial vascular image to obtain a brain tissue region vascular image.

[0022] Optionally, the step of performing blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image specifically includes:

[0023] The first image is filtered, and the filtered first image is then processed to extract blood vessels, resulting in a first blood vessel image.

[0024] The first image is subjected to feature enhancement processing, and the enhanced first image is subjected to blood vessel extraction processing to obtain a second blood vessel image;

[0025] The first vascular image and the second vascular image are fused to obtain an initial intracranial vascular image.

[0026] Optionally, the step of performing vessel segmentation processing on the groove region image to obtain a groove region vessel image specifically includes:

[0027] The image of the groove region is masked together with the initial intracranial vascular image to obtain a second image;

[0028] The vascular points in the groove region are determined based on the brain tissue region image and the initial intracranial vascular image.

[0029] The second image is searched based on the blood vessel points to obtain a blood vessel image of the groove region.

[0030] Optionally, the step of determining the vascular points in the groove region based on the brain tissue region image and the initial intracranial vascular image specifically includes:

[0031] A first set of points on the outer surface of the brain tissue region is obtained from the brain tissue region image;

[0032] Obtain the second point set of the initial intracranial vascular image;

[0033] The intersection of the first point set and the second point set is calculated to obtain the vascular points in the gully region.

[0034] Optionally, the step of searching the second image based on the blood vessel points to obtain the blood vessel image of the groove region specifically includes:

[0035] Using the blood vessel point as a seed point, search for the connected components where the seed point is located within the second image;

[0036] By merging the connected regions containing all seed points, a vascular image of the groove region is obtained.

[0037] A second aspect of the present invention provides a method for segmenting intracranial blood vessels, comprising the following steps:

[0038] Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image;

[0039] Brain tissue region images were determined based on the magnetic resonance angiography images and the modal images;

[0040] Based on the brain tissue region image, blood vessel segmentation processing of the brain tissue region is performed to obtain a blood vessel image of the brain tissue region.

[0041] Optionally, determining brain tissue region images based on the magnetic resonance angiography and the modal images specifically includes:

[0042] Remove non-brain tissue regions from the modal images;

[0043] The modal image excluding non-brain tissue regions is registered with the magnetic resonance angiography image to obtain a brain tissue region image.

[0044] Optionally, the step of performing blood vessel segmentation processing on the brain tissue region image to obtain a blood vessel image of the brain tissue region specifically includes:

[0045] Morphological processing was performed on the brain tissue region images to obtain intracranial images;

[0046] The intracranial image and the magnetic resonance angiography image are masked to obtain a first image;

[0047] The first image is processed by blood vessel segmentation to obtain an initial intracranial blood vessel image;

[0048] The brain tissue region image is masked with the initial intracranial vascular image to obtain a brain tissue region vascular image.

[0049] Optionally, the step of performing blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image specifically includes:

[0050] The first image is filtered, and the filtered first image is then processed to extract blood vessels, resulting in a first blood vessel image.

[0051] The first image is subjected to feature enhancement processing, and the enhanced first image is subjected to blood vessel extraction processing to obtain a second blood vessel image;

[0052] The first vascular image and the second vascular image are fused to obtain an initial intracranial vascular image.

[0053] A third aspect of the present invention provides a method for segmenting intracranial blood vessels, comprising the following steps:

[0054] Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image;

[0055] The groove region image is determined based on the magnetic resonance angiography image and the modal image;

[0056] Based on the image of the gully region, the blood vessels in the gully region are segmented to obtain the blood vessel image of the gully region.

[0057] Optionally, determining the groove region image based on the magnetic resonance vascular image and the modal image specifically includes:

[0058] Brain tissue region images were determined based on the magnetic resonance angiography images and the modal images;

[0059] Morphological processing was performed on the brain tissue region images to obtain intracranial images;

[0060] The sulcus region image is determined based on the brain tissue region image and the intracranial image.

[0061] Optionally, determining brain tissue region images based on the magnetic resonance angiography and the modal images specifically includes:

[0062] Remove non-brain tissue regions from the modal images;

[0063] The modal image excluding non-brain tissue regions is registered with the magnetic resonance angiography image to obtain a brain tissue region image.

[0064] Optionally, the step of performing vessel segmentation processing on the groove region image to obtain a groove region vessel image specifically includes:

[0065] The intracranial image and the magnetic resonance angiography image are masked to obtain a first image;

[0066] The first image is processed by blood vessel segmentation to obtain an initial intracranial blood vessel image;

[0067] The image of the groove region is masked together with the initial intracranial vascular image to obtain a second image;

[0068] The vascular points in the groove region are determined based on the brain tissue region image and the initial intracranial vascular image.

[0069] The second image is searched based on the blood vessel points to obtain a blood vessel image of the groove region.

[0070] Optionally, the step of performing blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image specifically includes:

[0071] The first image is filtered, and the filtered first image is then processed to extract blood vessels, resulting in a first blood vessel image.

[0072] The first image is subjected to feature enhancement processing, and the enhanced first image is subjected to blood vessel extraction processing to obtain a second blood vessel image;

[0073] The first vascular image and the second vascular image are fused to obtain an initial intracranial vascular image.

[0074] Optionally, the step of determining the vascular points in the groove region based on the brain tissue region image and the initial intracranial vascular image specifically includes:

[0075] A first set of points on the outer surface of the brain tissue region is obtained from the brain tissue region image;

[0076] Obtain the second point set of the initial intracranial vascular image;

[0077] The intersection of the first point set and the second point set is calculated to obtain the vascular points in the gully region.

[0078] Optionally, the step of searching the second image based on the blood vessel points to obtain the blood vessel image of the groove region specifically includes:

[0079] Using the blood vessel point as a seed point, search for the connected components where the seed point is located within the second image;

[0080] By merging the connected regions containing all seed points, a vascular image of the groove region is obtained.

[0081] A fourth aspect of the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method for segmenting intracranial blood vessels as described in the first, second, or third aspect.

[0082] A fifth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for segmenting intracranial blood vessels as described in the first, second, or third aspects.

[0083] The positive and progressive effects of this invention are as follows: by using magnetic resonance angiography images and corresponding modal images to determine brain tissue region images and sulcus region images respectively, and performing vascular segmentation processing on brain tissue region and sulcus region, the obtained brain tissue region vascular images and sulcus region vascular images are fused together to obtain a more complete and accurate intracranial vascular segmentation result, namely the target intracranial vascular image.

[0084] Furthermore, compared to determining brain tissue region images solely based on magnetic resonance angiography images, this invention determines brain tissue region images based on both magnetic resonance angiography images and modal images containing structural information. The brain tissue regions extracted from these images are more accurate. Furthermore, by performing vascular segmentation processing on the brain tissue region images, even more accurate vascular segmentation results for the brain tissue regions can be obtained.

[0085] Furthermore, compared to determining the groove region image solely based on magnetic resonance vascular images, this invention determines the groove region image based on both magnetic resonance vascular images and modal images containing structural information. The groove regions extracted from the groove region images are more accurate. Based on this, performing vascular segmentation processing on the groove region images can yield more accurate vascular segmentation results for the groove regions.

[0086] In addition, the intracranial blood vessel segmentation method provided by this invention does not rely on sample data and can be applied to various intracranial blood vessel segmentation scenarios, thus having a wider range of applications. Attached Figure Description

[0087] Figure 1 This is a flowchart of a method for segmenting intracranial blood vessels provided in Embodiment 1 of the present invention.

[0088] Figure 2 This is a schematic diagram of a modal image mask1 that removes non-brain tissue regions, provided as an embodiment of the present invention.

[0089] Figure 3 This is a flowchart of a method for segmenting blood vessels in a brain tissue region, as provided in Embodiment 1 of the present invention.

[0090] Figure 4 (a) is a schematic diagram of a first blood vessel image Vessel1 provided in an embodiment of the present invention.

[0091] Figure 4 (b) is a schematic diagram of a second blood vessel image, Vessel2, provided in an embodiment of the present invention.

[0092] Figure 4 (c) is a schematic diagram of an initial intracranial vascular image Vessel3 provided in an embodiment of the present invention.

[0093] Figure 4 (d) is a schematic diagram of a brain tissue region vascular image Vessel4 provided in an embodiment of the present invention.

[0094] Figure 5 This is a flowchart of a method for segmenting blood vessels in a groove region provided in Embodiment 1 of the present invention.

[0095] Figure 6 This is a schematic diagram of a second image provided in an embodiment of the present invention.

[0096] Figure 7 (a) is a schematic diagram of a blood vessel point 3 in a groove region provided in an embodiment of the present invention.

[0097] Figure 7 (b) is a schematic diagram of a vascular image Vessel5 in a groove region provided in an embodiment of the present invention.

[0098] Figure 8 This is a schematic diagram of a target intracranial blood vessel image provided in an embodiment of the present invention.

[0099] Figure 9 This is a structural block diagram of an intracranial blood vessel segmentation system provided in Embodiment 1 of the present invention.

[0100] Figure 10 This is a flowchart of a method for segmenting intracranial blood vessels provided in Embodiment 2 of the present invention.

[0101] Figure 11 This is a flowchart of a method for segmenting blood vessels in a brain tissue region, as provided in Embodiment 2 of the present invention.

[0102] Figure 12This is a structural block diagram of an intracranial blood vessel segmentation system provided in Embodiment 2 of the present invention.

[0103] Figure 13 This is a flowchart of a method for segmenting intracranial blood vessels provided in Embodiment 3 of the present invention.

[0104] Figure 14 This is a flowchart of a method for segmenting blood vessels in a groove region, as provided in Embodiment 3 of the present invention.

[0105] Figure 15 This is a structural block diagram of an intracranial blood vessel segmentation system provided in Embodiment 3 of the present invention.

[0106] Figure 16 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. Detailed Implementation

[0107] The present invention will be further illustrated by way of embodiments below, but the present invention is not limited to the scope of the embodiments described herein.

[0108] The intracranial region referred to in this invention is the interior of the skull, including the brain tissue region and the sulcus region. The brain tissue region includes white matter composed of nerve fibers or myelin sheaths. The sulcus region refers to the region that wraps around the brain tissue region, including many grooves or fissures of varying depths, as well as the sulci and gyri of the brain. The sulci and gyri refer to the ridges between the grooves or fissures.

[0109] Example 1

[0110] Figure 1 This is a flowchart illustrating a method for segmenting intracranial blood vessels provided in this embodiment. This method can be executed by an intracranial blood vessel segmentation system, which can be implemented through software and / or hardware. The intracranial blood vessel segmentation system can be part or all of an electronic device. In this embodiment, the electronic device can be a personal computer (PC), such as a desktop, all-in-one, laptop, or tablet computer, or it can be a mobile phone, wearable device, or PDA (Personal Digital Assistant) terminal device. The intracranial blood vessel segmentation method provided in this embodiment will be described below using an electronic device as the execution subject.

[0111] like Figure 1 As shown, the intracranial blood vessel segmentation method provided in this embodiment may include the following steps S1 to S5:

[0112] Step S1: Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image.

[0113] Magnetic resonance angiography (MRA) refers to images obtained using magnetic resonance angiography. MRA is a non-invasive vascular imaging method that does not require intubation or contrast agents. There are three main MRA vascular imaging data acquisition techniques: time-of-flight (TOF), phase contrast (PC), and contrast enhancement (CE). Specifically, MRA images acquired using TOF are called TOF-MRI, those acquired using PC are called PC-MRI, and those acquired using CE are called CE-MRI. The MRA images to be segmented can be TOF-MRI, PC-MRI, or CE-MRI, etc.

[0114] The modal image contains structural information and can also be called a structural item image. Specifically, it can be a structural item image of the T1 modality, a structural item image of the T2 modality, a structural item image of the FLAIR modality, or a structural item image of the T1ce modality, etc.

[0115] Specifically, a magnetic resonance imaging (MRI) device can be used to scan the head region of the target object to obtain the magnetic resonance vascular image to be segmented and the corresponding modal image. Alternatively, the magnetic resonance vascular image to be segmented and the corresponding modal image can be obtained from the network.

[0116] In specific implementations, both the magnetic resonance vascular image and the modal image can be two-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also two-dimensional images, with the corresponding image values ​​being pixel values. Alternatively, both the magnetic resonance vascular image and the modal image can be three-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also three-dimensional images, with the corresponding image values ​​being voxel values.

[0117] Step S2: Determine the brain tissue region image and sulcus region image based on the magnetic resonance angiography image and the modal image.

[0118] Specifically, brain tissue region images are determined based on the magnetic resonance angiography images and the modal images, and sulcus region images are determined based on the magnetic resonance angiography images and the modal images.

[0119] Specifically, the brain tissue region image and the sulcus region image are mask images. Specifically, in the brain tissue region image, the image value belonging to the brain tissue region is 1, and the image value not belonging to the brain tissue region is 0. Similarly, in the sulcus region image, the image value belonging to the sulcus region is 1, and the image value not belonging to the sulcus region is 0.

[0120] Step S3: Perform blood vessel segmentation processing on the brain tissue region image to obtain a blood vessel image of the brain tissue region.

[0121] Step S4: Perform blood vessel segmentation processing on the groove region image to obtain a blood vessel image of the groove region.

[0122] Step S5: Fuse the vascular images of the brain tissue region and the vascular images of the sulcus region to obtain the target intracranial vascular image. In specific implementations, the vascular vessels of the brain tissue region and the sulcus region can be observed through the target intracranial vascular image.

[0123] It should be noted that the above-mentioned blood vessel segmentation process can also be called blood vessel extraction process.

[0124] In the intracranial vessel segmentation method provided in this embodiment, brain tissue region images and sulcus region images are determined using magnetic resonance angiography (MRI) images and corresponding modal images, respectively. Vessel segmentation is then performed on both the brain tissue region and the sulcus region images. The resulting brain tissue region vessel images and sulcus region vessel images are then fused to obtain a more complete and accurate intracranial vessel segmentation result, i.e., the target intracranial vessel image. Furthermore, the intracranial vessel segmentation method provided in this embodiment does not rely on sample data and is applicable to various intracranial vessel segmentation scenarios, thus having a wider range of applications.

[0125] In one optional implementation, step S2 includes the following steps S21 to S22:

[0126] Step S21: Remove non-brain tissue regions from the modal image. These non-brain tissue regions may include dura mater material, external blood vessels, eyeglasses, fat, muscles, and the skull, etc. Removing non-brain tissue regions avoids interference from external information in the segmentation of intracranial blood vessels.

[0127] In a specific example, after removing non-brain tissue regions from the modal image, we can obtain, as follows: Figure 2 The modal image mask1 shows the region of non-brain tissue removed.

[0128] Step S22: Register the modal image mask1 (excluding non-brain tissue regions) with the magnetic resonance angiography image to obtain the brain tissue region image mask2.

[0129] Since the modal image corresponding to the magnetic resonance angiography image and the magnetic resonance angiography image have a certain positional offset in the voxel coordinate system, it is necessary to perform registration processing on the modal image mask1 that excludes non-brain tissue regions so that it corresponds to the magnetic resonance angiography image in the voxel coordinate system.

[0130] During the registration process, the magnetic resonance angiography image is used as a reference image, and the modal image corresponding to the magnetic resonance angiography image is used as a floating image. Specifically, the process includes the following steps:

[0131] First, a transformation matrix T is obtained from the modal image corresponding to the magnetic resonance angiography image to the magnetic resonance angiography image. The values ​​of each element in the transformation matrix T can be calculated based on the coordinates of corresponding points in the magnetic resonance angiography image and its corresponding modal image.

[0132] Secondly, the modal image mask1, which removes the non-brain tissue region, is transformed according to the transformation matrix T so that the modal image mask1, which removes the non-brain tissue region, maintains the same pose as the magnetic resonance angiography image in the voxel coordinate system, thereby obtaining the brain tissue region image mask2.

[0133] In this embodiment, by removing non-brain tissue regions and registering modal images of the removed non-brain tissue regions, the location of brain tissue regions in magnetic resonance angiography images can be effectively determined.

[0134] In one optional implementation, step S2 further includes steps S23 to S24:

[0135] Step S23: Perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3.

[0136] In specific implementation, the morphological processing can be dilation, opening operation, etc., to perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3, so as to ensure that the groove region is completely located inside the intracranial image mask3.

[0137] Step S24: Determine the sulcus region image mask4 based on the brain tissue region image mask2 and the intracranial image mask3. Specifically, the difference between the intracranial image mask3 and the brain tissue region image mask2 can be calculated to obtain the sulcus region image mask4.

[0138] Regarding step S3 above, this embodiment also provides a method for segmenting blood vessels in a brain tissue region, such as... Figure 3 As shown, it may specifically include the following steps S31 to S33:

[0139] Step S31: Mask the intracranial image mask3 with the magnetic resonance angiography image to obtain a first image. Specifically, the intracranial image mask3 and the magnetic resonance angiography image can be multiplied to obtain the first image. In the first image, image values ​​belonging to the intracranial region remain unchanged, that is, they are the same as the image values ​​at the corresponding positions in the magnetic resonance angiography image; image values ​​not belonging to the intracranial region are 0.

[0140] Step S32: Perform blood vessel segmentation processing on the first image to obtain the initial intracranial blood vessel image Vessel3.

[0141] In one alternative implementation, to improve the accuracy of the initial intracranial vascular image, such as... Figure 3 As shown, step S32 specifically includes the following steps S321 to S323:

[0142] Step S321: Filter the first image and perform blood vessel extraction processing on the filtered first image to obtain the first blood vessel image Vessel1. In a specific implementation, the filtering process can be a noise reduction process.

[0143] Step S322: Perform feature enhancement processing on the first image, and then perform blood vessel extraction processing on the feature-enhanced first image to obtain a second blood vessel image, Vessel2. In a specific implementation, the feature enhancement processing can be blood vessel enhancement processing.

[0144] Step S323: Fuse the first vascular image Vessel1 and the second vascular image Vessel2 to obtain the initial intracranial vascular image Vessel3.

[0145] Step S33: Mask the brain tissue region image mask2 with the initial intracranial vascular image Vessel3 to obtain the brain tissue region vascular image Vessel4. Specifically, the brain tissue region image mask2 can be multiplied with the initial intracranial vascular image Vessel3 to obtain the brain tissue region vascular image Vessel4. In Vessel4, the image values ​​of blood vessels belonging to the brain tissue region remain unchanged, that is, they are the same as the image values ​​at the corresponding positions in the initial intracranial vascular image Vessel3, while the image values ​​of blood vessels not belonging to the brain tissue region are 0.

[0146] In practice, to improve the accuracy of brain tissue region vascular images, noise reduction processing can be performed. For example, isolated blood vessels or small blood vessels can be removed from the brain tissue region vascular images to optimize the images and improve their visualization.

[0147] In a specific example, the first image is denoised, and then a threshold segmentation algorithm is used to extract blood vessels from the denoised first image, resulting in the following: Figure 4 The first blood vessel image Vessel1 is shown in (a). Blood vessel enhancement processing is performed on the first image, and then a threshold segmentation algorithm is used to extract blood vessels from the enhanced image, resulting in the following: Figure 4 (b) shows the second vascular image, Vessel2. After fusing the first vascular image, Vessel1, and the second vascular image, Vessel2, the following can be obtained: Figure 4 (c) shows the initial intracranial vascular image Vessel3. Masking the brain tissue region image mask2 with the initial intracranial vascular image Vessel3 yields the following result: Figure 4 (d) shows the Vessel4 image of the blood vessels in the brain tissue region, which is the result of blood vessel segmentation in the brain tissue region.

[0148] In addition to threshold segmentation algorithms, the methods for extracting blood vessels in steps S321 and S322 can also include statistical model segmentation algorithms, active contour segmentation algorithms, and hybrid model algorithms. Specifically, blood vessels are extracted using existing information in the first image, such as pixel grayscale information and shape information.

[0149] The vessels in the groove region are located in the groove between the skull and brain tissue, and they do not contact the skull. Because the skull size varies among different target subjects, it is impossible to accurately capture the groove region. To extract as many intracranial vascular images as possible, a larger groove region is usually selected; however, an excessively large groove region can introduce noise. To better extract the vessels in the groove region and effectively remove noise interference, a method for segmenting the vessels in the groove region in step S4 is proposed, such as... Figure 5 As shown, it may specifically include the following steps S41 to S43:

[0150] Step S41: Mask the groove region image mask4 and the initial intracranial vascular image Vessel3 to obtain a second image. The second image includes not only vascular points in the groove region but also noise points located near the skull.

[0151] In a specific example, the image mask4 of the gully region is compared with... Figure 4 (c) The initial intracranial vascular image Vessel3 shown can be masked to obtain the following: Figure 6The second image is shown. Specifically, the groove region image mask4 can be multiplied with the initial intracranial vascular image Vessel3 to obtain the second image.

[0152] Step S42: Determine the vascular points in the groove region based on the brain tissue region image mask2 and the initial intracranial vascular image Vessel3.

[0153] In one alternative implementation, such as Figure 5 As shown, step S42 specifically includes the following steps S421 to S423:

[0154] Step S421: Obtain the first point set point1 on the outer surface of the brain tissue region based on the brain tissue region image mask2.

[0155] Step S422: Obtain the second point set point2 of the initial intracranial vascular image Vessel3.

[0156] Step S423: Calculate the intersection of the first point set point1 and the second point set point2 to obtain the vascular point point3 in the groove region. Here, the vascular point point3 in the groove region is a point set, that is, a set composed of many points together. The vascular point point3 in the groove region is located both on the groove region and on the outer surface of the brain tissue region.

[0157] Step S43: Search the second image based on the blood vessel point 3 to obtain the vascular image Vessel 5 in the groove region.

[0158] In this embodiment, the second image is searched based on the blood vessel points in the groove region, which can remove noise points located near the skull in the second image, thereby obtaining the blood vessel image of the groove region, i.e., the blood vessel segmentation result of the groove region.

[0159] In one alternative implementation, such as Figure 5 As shown, step S43 above includes the following steps S431 to S432:

[0160] Step S431: Using the blood vessel point3 as the seed point, search for the connected component where the seed point is located in the second image.

[0161] Step S432: Merge the connected regions containing all seed points to obtain the vascular image Vessel5 of the groove region.

[0162] In this embodiment, the region where the blood vessel points are located in the gully region is obtained by searching the connected domain where the seed points are located, and the connected domains where all the seed points are located are merged. The regions in the gully region that are not in contact with the blood vessel points can be removed, thereby obtaining the blood vessel image of the gully region, i.e., the blood vessel segmentation result of the gully region.

[0163] In a specific example, based on the brain tissue region image mask2 and the initial intracranial vascular image Vessel3, the following can be obtained: Figure 7 (a) shows the blood vessel point 3 in the groove region. Searching the second image based on blood vessel point 3 yields the following results: Figure 7 (b) shows the vascular image of the groove region, Vessel5.

[0164] Furthermore, it will be like Figure 4 (d) shows the vascular image of the brain tissue region Vessel4 and Figure 7 (b) The Vessel5 images of the vascular region shown in the groove region can be merged to obtain the following: Figure 8 The image shown is of the target intracranial blood vessels.

[0165] like Figure 9 As shown, this embodiment also provides an intracranial blood vessel segmentation system 70, including an image acquisition module 71, an image determination module 72, a first segmentation module 73, a second segmentation module 74, and an image fusion module 75. The image acquisition module acquires a magnetic resonance imaging (MRI) image of the blood vessels to be segmented and a corresponding modal image. The image determination module determines a brain tissue region image and a sulcus region image based on the MRI image and the modal image. The first segmentation module performs blood vessel segmentation processing on the brain tissue region image to obtain a brain tissue region blood vessel image. The second segmentation module performs blood vessel segmentation processing on the sulcus region image to obtain a sulcus region blood vessel image. The image fusion module fuses the brain tissue region blood vessel image and the sulcus region blood vessel image to obtain a target intracranial blood vessel image.

[0166] In one optional implementation, the image determination module is specifically used to remove non-brain tissue regions from the modal image; and to register the modal image with the non-brain tissue regions removed with the magnetic resonance angiography image to obtain a brain tissue region image.

[0167] In one optional implementation, the image determination module is specifically used to perform morphological processing on the brain tissue region image to obtain an intracranial image; and to determine a sulcus region image based on the brain tissue region image and the intracranial image.

[0168] In one optional implementation, the first segmentation module includes a first masking unit, a blood vessel segmentation unit, and a second masking unit. The first masking unit performs masking processing on the intracranial image and the magnetic resonance angiography image to obtain a first image. The blood vessel segmentation unit performs blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image. The second masking unit performs masking processing on the brain tissue region image and the initial intracranial blood vessel image to obtain a brain tissue region blood vessel image.

[0169] In one optional embodiment, the blood vessel segmentation processing unit includes a first processing subunit, a second processing subunit, and an image fusion subunit. The first processing subunit performs filtering on the first image and blood vessel extraction on the filtered first image to obtain a first blood vessel image. The second processing subunit performs feature enhancement on the first image and blood vessel extraction on the feature-enhanced first image to obtain a second blood vessel image. The image fusion subunit fuses the first blood vessel image and the second blood vessel image to obtain an initial intracranial blood vessel image.

[0170] In one optional implementation, the second segmentation module includes a third masking unit, a vessel point determination unit, and an image search unit. The third masking unit performs masking processing on the groove region image and the initial intracranial vessel image to obtain a second image. The vessel point determination unit determines vessel points in the groove region based on the brain tissue region image and the initial intracranial vessel image. The image search unit searches the second image based on the vessel points to obtain a groove region vessel image.

[0171] In one optional implementation, the blood vessel point determination unit is specifically used to obtain a first set of points on the outer surface of the brain tissue region based on the brain tissue region image; obtain a second set of points on the initial intracranial blood vessel image; and calculate the intersection of the first set of points and the second set of points to obtain the blood vessel points in the groove region.

[0172] In one optional implementation, the image search unit is specifically used to search for the connected components where the blood vessel points are located in the second image, using the blood vessel points as seed points; and to merge the connected components where all seed points are located to obtain a blood vessel image of the groove region.

[0173] It should be noted that the intracranial blood vessel segmentation system in this embodiment can be a separate chip, chip module, or electronic device, or it can be a chip or chip module integrated into an electronic device.

[0174] The various modules / units included in the intracranial blood vessel segmentation system described in this embodiment can be software modules / units, hardware modules / units, or a combination of both.

[0175] Example 2

[0176] Figure 10 This is a flowchart illustrating a method for segmenting intracranial blood vessels according to this embodiment. This method can be executed by an intracranial blood vessel segmentation system, which can be implemented through software and / or hardware. The intracranial blood vessel segmentation system can be part or all of an electronic device. In this embodiment, the electronic device can be a personal computer (PC), such as a desktop, all-in-one, laptop, or tablet computer, or a mobile phone, wearable device, or PDA. The intracranial blood vessel segmentation method provided in this embodiment will be described below using an electronic device as the execution subject.

[0177] like Figure 10 As shown, the intracranial blood vessel segmentation method provided in this embodiment may include the following steps S101 to S103:

[0178] Step S101: Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image.

[0179] Magnetic resonance angiography (MRA) refers to images obtained using magnetic resonance angiography. MRA is a non-invasive vascular imaging method that does not require intubation or contrast agents. There are three main MRA vascular imaging data acquisition techniques: Time-of-Flight (TOF), Phase Contrast (PC), and Contrast Enhancement (CE). Magnetic resonance angiography images obtained using TOF acquisition are called TOF-MRI, those obtained using PC acquisition are called PC-MRI, and those obtained using CE acquisition are called CE-MRI. Specifically, the magnetic resonance angiography images to be segmented can be TOF-MRI, PC-MRI, or CE-MRI, etc.

[0180] The modal image contains structural information and can also be called a structural item image. Specifically, it can be a structural item image of the T1 modality, a structural item image of the T2 modality, a structural item image of the FLAIR modality, or a structural item image of the T1ce modality, etc.

[0181] Specifically, a magnetic resonance imaging (MRI) device can be used to scan the head region of the target object to obtain the magnetic resonance vascular image to be segmented and the corresponding modal image. Alternatively, the magnetic resonance vascular image to be segmented and the corresponding modal image can be obtained from the network.

[0182] In specific implementations, both the magnetic resonance vascular image and the modal image can be two-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also two-dimensional images, with the corresponding image values ​​being pixel values. Alternatively, both the magnetic resonance vascular image and the modal image can be three-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also three-dimensional images, with the corresponding image values ​​being voxel values.

[0183] Step S102: Determine the brain tissue region image based on the magnetic resonance angiography image and the modal image. Specifically, the brain tissue region image is a mask image. Specifically, in the brain tissue region image, image values ​​belonging to the brain tissue region are all 1, and image values ​​not belonging to the brain tissue region are all 0.

[0184] Step S103: Perform blood vessel segmentation processing on the brain tissue region image to obtain a blood vessel image of the brain tissue region. This blood vessel segmentation processing can also be called blood vessel extraction processing.

[0185] It should be noted that the segmentation result obtained by using the intracranial blood vessel segmentation method provided in this embodiment is a blood vessel image of the brain tissue region, that is, the blood vessel segmentation result of the brain tissue region.

[0186] Compared to determining brain tissue regions solely based on magnetic resonance angiography (MRI) images, this embodiment uses both MRI angiography images and modal images containing structural information to determine brain tissue regions. The extracted brain tissue regions are more accurate. Furthermore, performing vessel segmentation based on these brain tissue region images yields even more accurate vessel segmentation results. Additionally, the intracranial vessel segmentation method provided in this embodiment does not rely on sample data and is applicable to various intracranial vessel segmentation scenarios, thus having a wider range of applications.

[0187] In one optional embodiment, step S102 includes the following steps S102a to S102b:

[0188] Step S102a: Remove non-brain tissue regions from the modal image. These non-brain tissue regions may include dura mater material, external blood vessels, eyeglasses, fat, muscles, and the skull, etc. Removing non-brain tissue regions avoids interference from external information on intracranial blood vessel segmentation.

[0189] In a specific example, after removing non-brain tissue regions from the modal image, we can obtain, as follows: Figure 2 The modal image mask1 shows the region of non-brain tissue removed.

[0190] Step S102b: Register the modal image mask1, which excludes non-brain tissue regions, with the magnetic resonance angiography image to obtain the brain tissue region image mask2.

[0191] Since the modal image corresponding to the magnetic resonance angiography image and the magnetic resonance angiography image have a certain positional offset in the voxel coordinate system, it is necessary to perform registration processing on the modal image mask1 that excludes non-brain tissue regions so that it corresponds to the magnetic resonance angiography image in the voxel coordinate system.

[0192] During the registration process, the magnetic resonance angiography image is used as a reference image, and the modal image corresponding to the magnetic resonance angiography image is used as a floating image. Specifically, the process includes the following steps:

[0193] First, a transformation matrix T is obtained from the modal image corresponding to the magnetic resonance angiography image to the magnetic resonance angiography image. The values ​​of each element in the transformation matrix T can be calculated based on the coordinates of corresponding points in the magnetic resonance angiography image and its corresponding modal image.

[0194] Secondly, the modal image mask1, which removes the non-brain tissue region, is transformed according to the transformation matrix T so that the modal image mask1, which removes the non-brain tissue region, maintains the same pose as the magnetic resonance angiography image in the voxel coordinate system, thereby obtaining the brain tissue region image mask2.

[0195] In this embodiment, by removing non-brain tissue regions and registering modal images of the removed non-brain tissue regions, the location of brain tissue regions in magnetic resonance angiography images can be effectively determined.

[0196] In relation to step S103 above, this embodiment also provides a method for segmenting blood vessels in a brain tissue region, such as... Figure 11 As shown, the specific steps may include S103a to S103d:

[0197] Step S103a: Perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3.

[0198] In specific implementations, the morphological processing can include dilation, opening operations, etc., to perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3. In a specific example, the brain tissue region mask2 can be dilated to ensure that the groove regions are completely located inside the intracranial image mask3.

[0199] Step S103b: Mask the intracranial image mask3 with the magnetic resonance angiography image to obtain a first image. Specifically, the intracranial image mask3 and the magnetic resonance angiography image can be multiplied to obtain the first image. In the first image, image values ​​belonging to the intracranial region remain unchanged, that is, they are the same as the image values ​​at the corresponding positions in the magnetic resonance angiography image; image values ​​not belonging to the intracranial region are 0.

[0200] Step S103c: Perform blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image.

[0201] In one alternative implementation, to improve the accuracy of the initial intracranial vascular image, such as... Figure 11 As shown, the above step S103c specifically includes the following steps S103c1 to S103c3:

[0202] Step S103c1: Filter the first image and then perform blood vessel extraction processing on the filtered first image to obtain the first blood vessel image Vessel1. In a specific implementation, the filtering process can be a noise reduction process.

[0203] Step S103c2: Perform feature enhancement processing on the first image, and then perform blood vessel extraction processing on the feature-enhanced first image to obtain a second blood vessel image, Vessel2. In a specific implementation, the feature enhancement processing can be blood vessel enhancement processing.

[0204] Step S103c3: Fuse the first vascular image Vessel1 and the second vascular image Vessel2 to obtain the initial intracranial vascular image Vessel3.

[0205] Step S103d: Mask the brain tissue region image and the initial intracranial vascular image to obtain a brain tissue region vascular image.

[0206] The brain tissue region image mask2 is masked together with the initial intracranial vascular image Vessel3 to obtain the brain tissue region vascular image Vessel4. Specifically, the brain tissue region image mask2 can be multiplied by the initial intracranial vascular image Vessel3 to obtain the brain tissue region vascular image Vessel4. In Vessel4, the image values ​​of vessels belonging to the brain tissue region remain unchanged, that is, they are the same as the image values ​​of the corresponding positions in the initial intracranial vascular image Vessel3, while the image values ​​of vessels not belonging to the brain tissue region are 0.

[0207] In practice, to improve the accuracy of brain tissue region vascular images, noise reduction processing can be performed. For example, isolated blood vessels or small blood vessels can be removed from the brain tissue region vascular images to optimize the images and improve their visualization.

[0208] In a specific example, the first image is denoised, and then a threshold segmentation algorithm is used to extract blood vessels from the denoised first image, resulting in the following: Figure 4 The first blood vessel image Vessel1 is shown in (a). Blood vessel enhancement processing is performed on the first image, and then a threshold segmentation algorithm is used to extract blood vessels from the enhanced image, resulting in the following: Figure 4 (b) shows the second vascular image, Vessel2. After fusing the first vascular image, Vessel1, and the second vascular image, Vessel2, the following can be obtained: Figure 4 (c) shows the initial intracranial vascular image Vessel3. Masking the brain tissue region image mask2 with the initial intracranial vascular image Vessel3 yields the following result: Figure 4 (d) shows the Vessel4 image of the blood vessels in the brain tissue region, which is the result of blood vessel segmentation in the brain tissue region.

[0209] In addition to threshold segmentation algorithms, the methods for extracting blood vessels in steps S103c1 and S103c2 can also include statistical model segmentation algorithms, active contour segmentation algorithms, and hybrid model algorithms. Specifically, blood vessels are extracted using existing information in the first image, such as pixel grayscale information and shape information.

[0210] like Figure 12 As shown, this embodiment also provides an intracranial blood vessel segmentation system 80, including an image acquisition module 81, an image determination module 82, and a blood vessel segmentation module 83. The image acquisition module is used to acquire a magnetic resonance imaging (MRI) image of the blood vessel to be segmented and a corresponding modal image. The image determination module is used to determine a brain tissue region image based on the MRI image of the blood vessel and the modal image. The blood vessel segmentation module is used to perform blood vessel segmentation processing on the brain tissue region image to obtain a blood vessel image of the brain tissue region.

[0211] In one optional implementation, the image determination module is specifically used to remove non-brain tissue regions from the modal image; and to register the modal image with the non-brain tissue regions removed with the magnetic resonance angiography image to obtain a brain tissue region image.

[0212] In one optional embodiment, the blood vessel segmentation module includes a morphological processing unit, a first masking unit, a blood vessel segmentation unit, and a second masking unit. The morphological processing unit performs morphological processing on the brain tissue region image to obtain an intracranial image. The first masking unit performs masking processing on the intracranial image and the magnetic resonance angiography image to obtain a first image. The blood vessel segmentation unit performs blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image. The second masking unit performs masking processing on the brain tissue region image and the initial intracranial blood vessel image to obtain a brain tissue region blood vessel image.

[0213] In one optional embodiment, the vessel segmentation unit includes a first processing subunit, a second processing subunit, and an image fusion subunit. The first processing subunit filters the first image and performs vessel extraction processing on the filtered first image to obtain a first vessel image. The second processing subunit enhances the first image and performs vessel extraction processing on the feature-enhanced first image to obtain a second vessel image. The image fusion subunit fuses the first vessel image and the second vessel image to obtain an initial intracranial vessel image.

[0214] It should be noted that the intracranial blood vessel segmentation system in this embodiment can be a separate chip, chip module, or electronic device, or it can be a chip or chip module integrated into an electronic device.

[0215] The various modules / units included in the intracranial blood vessel segmentation system described in this embodiment can be software modules / units, hardware modules / units, or a combination of both.

[0216] Example 3

[0217] Figure 13 This is a flowchart illustrating a method for segmenting intracranial blood vessels according to this embodiment. This method can be executed by an intracranial blood vessel segmentation system, which can be implemented through software and / or hardware. The intracranial blood vessel segmentation system can be part or all of an electronic device. In this embodiment, the electronic device can be a personal computer (PC), such as a desktop, all-in-one, laptop, or tablet computer, or a mobile phone, wearable device, or PDA. The intracranial blood vessel segmentation method provided in this embodiment will be described below using an electronic device as the execution subject.

[0218] like Figure 13 As shown, the intracranial blood vessel segmentation method provided in this embodiment may include the following steps S201 to S203:

[0219] Step S201: Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image.

[0220] Magnetic resonance angiography (MRA) refers to images obtained using magnetic resonance angiography. MRA is a non-invasive vascular imaging method that does not require intubation or contrast agents. There are three main MRA vascular imaging data acquisition techniques: Time-of-Flight (TOF), Phase Contrast (PC), and Contrast Enhancement (CE). Magnetic resonance angiography images obtained using TOF acquisition are called TOF-MRI, those obtained using PC acquisition are called PC-MRI, and those obtained using CE acquisition are called CE-MRI. Specifically, the magnetic resonance angiography images to be segmented can be TOF-MRI, PC-MRI, or CE-MRI, etc.

[0221] The modal image contains structural information and can also be called a structural item image. Specifically, it can be a structural item image of the T1 modality, a structural item image of the T2 modality, a structural item image of the FLAIR modality, or a structural item image of the T1ce modality, etc.

[0222] Specifically, a magnetic resonance imaging (MRI) device can be used to scan the head region of the target object to obtain the magnetic resonance vascular image to be segmented and the corresponding modal image. Alternatively, the magnetic resonance vascular image to be segmented and the corresponding modal image can be obtained from the network.

[0223] In specific implementations, both the magnetic resonance vascular image and the modal image can be two-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also two-dimensional images, with the corresponding image values ​​being pixel values. Alternatively, both the magnetic resonance vascular image and the modal image can be three-dimensional images. Correspondingly, the images determined based on the magnetic resonance vascular image and the modal image are also three-dimensional images, with the corresponding image values ​​being voxel values.

[0224] Step S202: Determine the groove region image based on the magnetic resonance vascular image and the modal image.

[0225] Specifically, the gully region image is a mask image. In particular, the image value belonging to the gully region is 1, and the image value not belonging to the gully region is 0.

[0226] Step S203: Perform vessel segmentation processing on the groove region image to obtain a groove region vessel image. This vessel segmentation processing can also be referred to as vessel extraction processing.

[0227] It should be noted that the segmentation result obtained by using the intracranial blood vessel segmentation method provided in this embodiment is a groove region blood vessel image, that is, the blood vessel segmentation result of the groove region.

[0228] Compared to determining the groove region image solely based on magnetic resonance angiography (MRI) images, this embodiment determines the groove region image by combining MRI images with modal images containing structural information. The groove regions extracted from the groove region image are more accurate. Based on this, vessel segmentation processing is performed on the groove region image, resulting in more accurate vessel segmentation results for the groove region. Furthermore, the intracranial vessel segmentation method provided in this embodiment does not rely on sample data and is applicable to various intracranial vessel segmentation scenarios, thus having a wider range of applications.

[0229] In one optional implementation, step S202 includes the following steps S202a to S202c:

[0230] Step S202a: Determine brain tissue region images based on the magnetic resonance angiography and the modal images.

[0231] Step S202a may include the following steps S202a1 to S202a2:

[0232] Step S202a1: Remove non-brain tissue regions from the modal image. These non-brain tissue regions may include dura mater material, external blood vessels, eyeglasses, fat, muscles, and the skull, etc. Removing non-brain tissue regions avoids interference from external information on intracranial blood vessel segmentation.

[0233] In a specific example, after removing non-brain tissue regions from the modal image, we can obtain, as follows: Figure 2 The modal image mask1 shows the region of non-brain tissue removed.

[0234] Step S202a2: Register the modal image mask1 (excluding non-brain tissue regions) with the magnetic resonance angiography image to obtain the brain tissue region image mask2.

[0235] Since the modal image corresponding to the magnetic resonance angiography image and the magnetic resonance angiography image have a certain positional offset in the voxel coordinate system, it is necessary to perform registration processing on the modal image mask1 that excludes non-brain tissue regions so that it corresponds to the magnetic resonance angiography image in the voxel coordinate system.

[0236] During the registration process, the magnetic resonance angiography image is used as a reference image, and the modal image corresponding to the magnetic resonance angiography image is used as a floating image. Specifically, the process includes the following steps:

[0237] First, a transformation matrix T is obtained from the modal image corresponding to the magnetic resonance angiography image to the magnetic resonance angiography image. The values ​​of each element in the transformation matrix T can be calculated based on the coordinates of corresponding points in the magnetic resonance angiography image and its corresponding modal image.

[0238] Secondly, the modal image mask1, which removes the non-brain tissue region, is transformed according to the transformation matrix T so that the modal image mask1, which removes the non-brain tissue region, maintains the same pose as the magnetic resonance angiography image in the voxel coordinate system, thereby obtaining the brain tissue region image mask2.

[0239] In this embodiment, by removing non-brain tissue regions and registering modal images of the removed non-brain tissue regions, the location of brain tissue regions in magnetic resonance angiography images can be effectively determined.

[0240] Step S202b: Perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3.

[0241] In specific implementation, the morphological processing can be dilation, opening operation, etc., to perform morphological processing on the brain tissue region image mask2 to obtain the intracranial image mask3, so as to ensure that the groove region is completely located inside the intracranial image mask3.

[0242] Step S202c: Determine the sulcus region image mask4 based on the brain tissue region image mask2 and the intracranial image mask3. Specifically, the difference between the intracranial image mask3 and the brain tissue region image mask2 can be calculated to obtain the sulcus region image mask4.

[0243] The vessels in the groove region are located in the groove region between the skull and brain tissue, and the vessels in the groove region do not contact the skull. Because the skull size varies among different target subjects, it is impossible to accurately capture the groove region. To extract as many intracranial vascular images as possible, a larger groove region is usually selected; however, an excessively large groove region can introduce noise. To better extract the vessels in the groove region and effectively remove noise interference, a vessel segmentation method for the groove region in step S203 is proposed, such as... Figure 14 As shown, the specific steps may include S203a to S203e:

[0244] Step S203a: Mask the intracranial image mask3 with the magnetic resonance angiography image to obtain a first image. Specifically, the intracranial image mask3 can be multiplied with the magnetic resonance angiography image to obtain the first image. In the first image, image values ​​belonging to the intracranial region remain unchanged, that is, they are the same as the image values ​​at the corresponding positions in the magnetic resonance angiography image; image values ​​not belonging to the intracranial region are 0.

[0245] Step S203b: Perform blood vessel segmentation processing on the first image to obtain the initial intracranial blood vessel image Vessel3.

[0246] In one alternative implementation, to improve the accuracy of the initial intracranial vascular image, such as... Figure 14 As shown, step S203b specifically includes the following steps S203b1 to S203b3:

[0247] Step S203b1: Filter the first image and then perform blood vessel extraction processing on the filtered first image to obtain the first blood vessel image Vessel1. In a specific implementation, the filtering process can be a noise reduction process.

[0248] Step S203b2: Perform feature enhancement processing on the first image, and then perform blood vessel extraction processing on the feature-enhanced first image to obtain a second blood vessel image, Vessel2. In a specific implementation, the feature enhancement processing can be blood vessel enhancement processing.

[0249] Step S203b3: Fuse the first vascular image Vessel1 and the second vascular image Vessel2 to obtain the initial intracranial vascular image Vessel3.

[0250] Step S203c: Mask the groove region image mask4 and the initial intracranial vascular image Vessel3 to obtain a second image. The second image includes not only vascular points in the groove region but also noise points located near the skull.

[0251] In a specific example, the image mask4 of the gully region is compared with... Figure 4 (c) The initial intracranial vascular image Vessel3 shown can be masked to obtain the following: Figure 6 The second image is shown. Specifically, the groove region image mask4 can be multiplied with the initial intracranial vascular image Vessel3 to obtain the second image.

[0252] Step S203d: Determine the vascular points in the groove region based on the brain tissue region image mask2 and the initial intracranial vascular image Vessel3.

[0253] In one alternative implementation, such as Figure 14 As shown, the above step S203d specifically includes the following steps S203d1 to S203d3:

[0254] Step S203d1: Obtain the first point set point1 on the outer surface of the brain tissue region based on the brain tissue region image mask2.

[0255] Step S203d2: Obtain the second point set point2 of the initial intracranial vascular image Vessel3.

[0256] Step S203d3: Calculate the intersection of the first point set point1 and the second point set point2 to obtain the vascular point point3 in the sulcus region. Here, the vascular point point3 in the sulcus region is a point set, that is, a set composed of many points together. The vascular point point3 in the sulcus region is located both on the sulcus region and on the outer surface of the brain tissue region.

[0257] Step S203e: Search the second image based on the blood vessel points to obtain a blood vessel image of the groove region.

[0258] In this embodiment, the second image is searched based on the blood vessel points in the groove region, which can remove noise points located near the skull in the second image, thereby obtaining the blood vessel image of the groove region, i.e., the blood vessel segmentation result of the groove region.

[0259] In one alternative implementation, such as Figure 14 As shown, the above step S203e includes the following steps SS203e1 to SS203e2:

[0260] Step S203e1: Using the blood vessel point3 as the seed point, search for the connected component where the seed point is located in the second image.

[0261] Step S203e2: Merge the connected regions containing all seed points to obtain the vascular image Vessel5 of the groove region.

[0262] In this embodiment, the region where the blood vessel points are located in the gully region is obtained by searching the connected domain where the seed points are located, and the connected domains where all the seed points are located are merged. The regions in the gully region that are not in contact with the blood vessel points can be removed, thereby obtaining the blood vessel image of the gully region, i.e., the blood vessel segmentation result of the gully region.

[0263] In a specific example, based on the brain tissue region image mask2 and the initial intracranial vascular image Vessel3, the following can be obtained: Figure 7(a) shows the blood vessel point 3 in the groove region. Searching the second image based on blood vessel point 3 yields the following results: Figure 7 (b) shows the vascular image of the groove region, Vessel5.

[0264] Furthermore, it will be like Figure 4 (d) shows the vascular image of the brain tissue region Vessel4 and Figure 7 (b) The Vessel5 images of the vascular region shown in the groove region can be merged to obtain the following: Figure 8 The image shown is of the target intracranial blood vessels.

[0265] like Figure 15 As shown, this embodiment also provides an intracranial blood vessel segmentation system 90, including an image acquisition module 91, an image determination module 92, and a blood vessel segmentation module 93. The image acquisition module acquires a magnetic resonance imaging (MRI) image of the blood vessel to be segmented and a corresponding modal image. The image determination module determines a groove region image based on the MRI image and the modal image. The blood vessel segmentation module performs blood vessel segmentation processing on the groove region image to obtain a groove region blood vessel image.

[0266] In one optional implementation, the image determination module includes a first determination unit, a morphological processing unit, and a second determination unit. The first determination unit is used to determine a brain tissue region image based on the magnetic resonance angiography image and the modal image. The morphological processing unit is used to perform morphological processing on the brain tissue region image to obtain an intracranial image. The second determination unit is used to determine a sulcus region image based on the brain tissue region image and the intracranial image.

[0267] In one optional implementation, the first determining unit is specifically used to remove non-brain tissue regions from the modal image; and to register the modal image with the non-brain tissue regions removed with the magnetic resonance angiography image to obtain a brain tissue region image.

[0268] In one optional embodiment, the vessel segmentation module includes a first masking unit, a vessel segmentation unit, a second masking unit, a vessel point determination unit, and an image search unit. The first masking unit performs masking processing on the intracranial image and the magnetic resonance angiography image to obtain a first image. The vessel segmentation unit performs vessel segmentation processing on the first image to obtain an initial intracranial vessel image. The second masking unit performs masking processing on the groove region image and the initial intracranial vessel image to obtain a second image. The vessel point determination unit determines vessel points in the groove region based on the brain tissue region image and the initial intracranial vessel image. The image search unit searches the second image based on the vessel points to obtain a groove region vessel image.

[0269] In one optional embodiment, the vessel segmentation unit includes a first processing subunit, a second processing subunit, and an image fusion subunit. The first processing subunit filters the first image and performs vessel extraction processing on the filtered first image to obtain a first vessel image. The second processing subunit enhances the first image and performs vessel extraction processing on the feature-enhanced first image to obtain a second vessel image. The image fusion subunit fuses the first vessel image and the second vessel image to obtain an initial intracranial vessel image.

[0270] In one optional implementation, the blood vessel point determination unit is specifically used to obtain a first set of points on the outer surface of the brain tissue region based on the brain tissue region image; obtain a second set of points on the initial intracranial blood vessel image; and calculate the intersection of the first set of points and the second set of points to obtain the blood vessel points in the groove region.

[0271] In one optional implementation, the image search unit is specifically used to search for the connected components where the blood vessel points are located in the second image, using the blood vessel points as seed points; and to merge the connected components where all seed points are located to obtain a trough region blood vessel image.

[0272] It should be noted that the intracranial blood vessel segmentation system in this embodiment can be a separate chip, chip module, or electronic device, or it can be a chip or chip module integrated into an electronic device.

[0273] The various modules / units included in the intracranial blood vessel segmentation system described in this embodiment can be software modules / units, hardware modules / units, or a combination of both.

[0274] Example 4

[0275] Figure 16 This is a schematic diagram of the structure of an electronic device provided in this embodiment. The electronic device includes at least one processor and a memory communicatively connected to the at least one processor. The memory stores a computer program that can be executed by the at least one processor, enabling the at least one processor to perform the intracranial blood vessel segmentation method of Embodiments 1, 2, or 3. The electronic device provided in this embodiment can be a personal computer, such as a desktop computer, all-in-one computer, laptop computer, tablet computer, etc., or it can be a mobile phone, wearable device, PDA, or other terminal device. Figure 16 The electronic device 3 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0276] The components of the electronic device 3 may include, but are not limited to: at least one processor 4, at least one memory 5, and a bus 6 connecting different system components (including memory 5 and processor 4).

[0277] Bus 6 includes a data bus, an address bus, and a control bus.

[0278] The memory 5 may include volatile memory, such as random access memory (RAM) 51 and / or cache memory 52, and may further include read-only memory (ROM) 53.

[0279] The memory 5 may also include a program / utility 55 having a set (at least one) of program modules 54, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0280] The processor 4 executes various functional applications and data processing by running computer programs stored in the memory 5, such as the intracranial blood vessel segmentation method of embodiments 1, 2 or 3 above.

[0281] Electronic device 3 can also communicate with one or more external devices 7 (e.g., keyboard, pointing device, etc.). This communication can be performed through input / output (I / O) interface 8. Furthermore, electronic device 3 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 9. Figure 16 As shown, network adapter 9 communicates with other modules of electronic device 3 via bus 6. It should be understood that, although... Figure 16 Not shown, it can be combined with electronic device 3 to use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0282] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0283] Example 5

[0284] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the intracranial blood vessel segmentation method of Embodiment 1, 2, or 3.

[0285] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0286] In a possible implementation, the present invention can also be implemented as a program product comprising program code that, when the program product is run on an electronic device, causes the electronic device to perform the intracranial vessel segmentation method of embodiment 1, 2, or 3.

[0287] The program code for executing the present invention can be written in any combination of one or more programming languages. The program code can be executed entirely on an electronic device, partially on an electronic device, as a standalone software package, partially on an electronic device and partially on a remote device, or entirely on a remote device.

[0288] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but all such changes and modifications fall within the scope of protection of the present invention.

Claims

1. A method for segmenting intracranial blood vessels, characterized in that, Includes the following steps: Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image; Brain tissue region images are determined based on the magnetic resonance angiography images and the modal images. Morphological processing is performed on the brain tissue region images to obtain intracranial images. Sulcus region images are determined based on the brain tissue region images and the intracranial images. The brain tissue region image is processed to segment blood vessels to obtain a brain tissue region blood vessel image. Specifically, this includes: masking the intracranial image and the magnetic resonance angiography image to obtain a first image; performing blood vessel segmentation on the first image to obtain an initial intracranial blood vessel image; and masking the brain tissue region image and the initial intracranial blood vessel image to obtain a brain tissue region blood vessel image. The process of segmenting blood vessels in the groove region based on the groove region image to obtain a groove region blood vessel image specifically includes: masking the groove region image and the initial intracranial blood vessel image to obtain a second image; obtaining a first set of points on the outer surface of the brain tissue region based on the brain tissue region image; obtaining a second set of points on the initial intracranial blood vessel image; calculating the intersection of the first set of points and the second set of points to obtain blood vessel points in the groove region; and searching the second image based on the blood vessel points to obtain the groove region blood vessel image. By fusing the vascular images of the brain tissue region and the vascular images of the groove region, a target intracranial vascular image is obtained.

2. The segmentation method as described in claim 1, characterized in that, Determining brain tissue region images based on the magnetic resonance angiography images and the modal images specifically includes: Remove non-brain tissue regions from the modal images; The modal image excluding non-brain tissue regions is registered with the magnetic resonance angiography image to obtain a brain tissue region image.

3. The segmentation method as described in claim 1, characterized in that, The step of performing blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image specifically includes: The first image is filtered, and the filtered first image is then processed to extract blood vessels, resulting in a first blood vessel image. The first image is subjected to feature enhancement processing, and the enhanced first image is subjected to blood vessel extraction processing to obtain a second blood vessel image; The first vascular image and the second vascular image are fused to obtain an initial intracranial vascular image.

4. The segmentation method as described in claim 1, characterized in that, The step of searching the second image based on the blood vessel points to obtain the blood vessel image of the groove region specifically includes: Using the blood vessel point as a seed point, search for the connected components where the seed point is located within the second image; By merging the connected regions containing all seed points, a vascular image of the groove region is obtained.

5. A method for segmenting intracranial blood vessels, characterized in that, Includes the following steps: Obtain the magnetic resonance vascular image to be segmented and the corresponding modal image; Determining the sulcus region image based on the magnetic resonance angiography image and the modal image specifically includes: determining the brain tissue region image based on the magnetic resonance angiography image and the modal image; performing morphological processing on the brain tissue region image to obtain an intracranial image; and determining the sulcus region image based on the brain tissue region image and the intracranial image. The process of segmenting blood vessels in the groove region based on the groove region image to obtain a groove region blood vessel image specifically includes: masking the intracranial image and the magnetic resonance angiography image to obtain a first image; performing blood vessel segmentation on the first image to obtain an initial intracranial blood vessel image; masking the groove region image and the initial intracranial blood vessel image to obtain a second image; obtaining a first set of points on the outer surface of the brain tissue region based on the brain tissue region image; obtaining a second set of points on the initial intracranial blood vessel image; calculating the intersection of the first set of points and the second set of points to obtain blood vessel points in the groove region; and searching the second image based on the blood vessel points to obtain the groove region blood vessel image.

6. The segmentation method as described in claim 5, characterized in that, Determining brain tissue region images based on the magnetic resonance angiography images and the modal images specifically includes: Remove non-brain tissue regions from the modal images; The modal image excluding non-brain tissue regions is registered with the magnetic resonance angiography image to obtain a brain tissue region image.

7. The segmentation method as described in claim 5, characterized in that, The step of performing blood vessel segmentation processing on the first image to obtain an initial intracranial blood vessel image specifically includes: The first image is filtered, and the filtered first image is then processed to extract blood vessels, resulting in a first blood vessel image. The first image is subjected to feature enhancement processing, and the enhanced first image is subjected to blood vessel extraction processing to obtain a second blood vessel image; The first vascular image and the second vascular image are fused to obtain an initial intracranial vascular image.

8. The segmentation method as described in claim 5, characterized in that, The step of searching the second image based on the blood vessel points to obtain the blood vessel image of the groove region specifically includes: Using the blood vessel point as a seed point, search for the connected components where the seed point is located within the second image; By merging the connected regions containing all seed points, a vascular image of the groove region is obtained.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for segmenting intracranial blood vessels as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for segmenting intracranial blood vessels as described in any one of claims 1-8.

Images