Medical image segmentation method and device, computer equipment and storage medium

Abstract

The invention relates to a medical image segmentation method and device, computer equipment and a storage medium. The method comprises the following steps: performing morphological processing on a mask image corresponding to an original image of a target part to obtain a morphological image of the target part; the morphological image comprises an expansion image and a corrosion image; determining a surface area of the target part from the original image according to the expansion image and the corrosion image, and determining an internal area of the target part from the original image according to the corrosion image; determining a surface segmentation result of a blood vessel of the target part in the surface area, and determining an internal segmentation result of the blood vessel in the internal area; and obtaining a target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result. By adopting the method, the segmentation precision of the target part can be improved, and accurate extraction of the target part is realized.

Description

Technical Field

[0001] The present application relates to the technical field of medical image processing, and in particular to a medical image segmentation method, apparatus, computer equipment and storage medium. Background Art

[0002] Neurosurgery requires precise positioning of lesions before surgery, design and optimization of surgical pathways, precise resection of lesion areas during surgery, and postoperative evaluation of lesion resection effects and disease recovery. Accurate extraction of cerebral blood vessels is crucial in neurosurgery, as it helps to better distinguish blood vessels from surrounding tissues, observe blood vessels more clearly, and provide more accurate information for neurosurgery.

[0003] Cranial CTA (Computed Tomography Angiography) is a technology that uses CT images to perform three-dimensional reconstruction of cranial and cerebral blood vessels. It can be used to assist neurosurgery. However, in addition to blood vessels, CTA also contains non-vascular structures such as white matter, gray matter, tissue fluid, bones, and air gaps. These structures are located relatively closely together. Using traditional threshold segmentation, multi-scale spatial filtering, statistical model segmentation, active contour segmentation, and hybrid model methods, it is difficult to accurately extract the vascular structure in CTA due to the low segmentation accuracy.

[0004] Therefore, the current medical image segmentation technology for CTA has the problem of low segmentation accuracy. Summary of the invention

[0005] Based on this, it is necessary to provide a medical image segmentation method, apparatus, computer equipment, computer-readable storage medium and computer program product with high segmentation accuracy in order to address the above technical problems.

[0006] In a first aspect, the present application provides a medical image segmentation method. The method comprises:

[0007] Performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0008] Determining a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining an internal area of ​​the target part from the original image according to the eroded image;

[0009] Determine a surface segmentation result of the blood vessel at the target site in the surface area, and determine an internal segmentation result of the blood vessel in the internal area;

[0010] A target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result.

[0011] In one embodiment, determining the surface segmentation result of the blood vessel at the target site in the surface area includes:

[0012] Performing segmentation processing on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area;

[0013] removing the first target from the first segmentation result to obtain a second segmentation result of the surface area;

[0014] determining a second target from the second segmentation result, performing segmentation processing on the second target, and obtaining a third segmentation result of the surface area;

[0015] The second segmentation result and the third segmentation result are combined to obtain the surface segmentation result.

[0016] In one embodiment, the segmentation process is performed on the surface area according to the preset segmentation model to obtain the first segmentation result of the surface area, including:

[0017] Performing enhancement processing on the blood vessels in the surface area to obtain an enhanced surface area;

[0018] The enhanced surface area is segmented according to the segmentation model to obtain the first segmentation result.

[0019] In one of the embodiments, before removing the first target from the first segmentation result to obtain the second segmentation result of the surface area, the method further includes:

[0020] Determining a connected domain in the first segmentation result;

[0021] When the property of the connected domain meets a first preset condition, the connected domain is determined as the first target; the property includes at least one of shape and volume, and the first preset condition includes at least one of abnormal shape and abnormal volume.

[0022] In one embodiment, determining the second target from the second segmentation result, performing segmentation processing on the second target, and obtaining the third segmentation result of the surface area includes:

[0023] Determining a maximum connected domain in the second segmentation result;

[0024] Performing expansion processing on the maximum connected domain to obtain the second target;

[0025] The blood vessel in the second target is segmented to obtain the third segmentation result.

[0026] In one embodiment, determining the surface area of ​​the target part from the original image according to the dilated image and the eroded image includes:

[0027] Obtaining a surface mask image of the target part according to the intersection of the dilation image and the erosion image;

[0028] The original image is subjected to mask processing according to the surface mask image to obtain the surface area of ​​the target part.

[0029] In one embodiment, determining the internal area of ​​the target part from the original image according to the eroded image includes:

[0030] Obtaining an internal mask image of the target part according to the corrosion image;

[0031] The original image is subjected to mask processing according to the internal mask image to obtain the internal area of ​​the target part.

[0032] In a second aspect, the present application also provides a medical image segmentation device. The device comprises:

[0033] A processing module, used for performing morphological processing on a mask image corresponding to an original image of a target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0034] a determination module, configured to determine a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and to determine an internal area of ​​the target part from the original image according to the eroded image;

[0035] a segmentation module, used for determining a surface segmentation result of the blood vessel at the target site in the surface area, and determining an internal segmentation result of the blood vessel in the internal area;

[0036] A merging module is used to obtain a target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result.

[0037] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0038] Performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0039] Determining a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining an internal area of ​​the target part from the original image according to the eroded image;

[0040] Determine a surface segmentation result of the blood vessel at the target site in the surface area, and determine an internal segmentation result of the blood vessel in the internal area;

[0041] A target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result.

[0042] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0043] Performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0044] Determining a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining an internal area of ​​the target part from the original image according to the eroded image;

[0045] Determine a surface segmentation result of the blood vessel at the target site in the surface area, and determine an internal segmentation result of the blood vessel in the internal area;

[0046] A target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result.

[0047] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0048] Performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0049] Determining a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining an internal area of ​​the target part from the original image according to the eroded image;

[0050] Determine a surface segmentation result of the blood vessel at the target site in the surface area, and determine an internal segmentation result of the blood vessel in the internal area;

[0051] A target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result.

[0052] The above-mentioned medical image segmentation method, device, computer equipment, storage medium and computer program product obtain a morphological image of the target part by morphologically processing a mask image corresponding to the original image of the target part, and the morphological image includes a dilation image and an erosion image. According to the dilation image and the erosion image, the surface area of ​​the target part is determined from the original image, and according to the erosion image, the internal area of ​​the target part is determined from the original image, and the surface segmentation result of the blood vessels in the surface area of ​​the target part is determined, and the internal segmentation result of the blood vessels in the internal area is determined. According to the surface segmentation result and the internal segmentation result, the target segmentation result of the blood vessels is obtained; the target part can be divided into a surface area and an internal area in the medical image, and different image segmentation processing is performed respectively according to the characteristics of the surface area and the internal area to improve the respective segmentation accuracy, and then by merging the surface segmentation result and the internal segmentation result, the segmentation accuracy of the entire target part is improved, and the target part is accurately extracted. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a schematic flow chart of a medical image segmentation method in one embodiment;

[0054] Figure 2 is a schematic diagram of a flow chart of a brain tissue region blood vessel segmentation process in one embodiment;

[0055] Figure 3 A schematic diagram of a flow chart of a blood vessel segmentation step in an internal area of ​​brain tissue in an embodiment;

[0056] Figure 4 A schematic diagram of a flow chart of a blood vessel segmentation step in a brain tissue surface area in an embodiment;

[0057] Figure 5 is a schematic diagram of a surface area of ​​brain tissue in one embodiment;

[0058] Figure 6 is a schematic diagram of an internal region of brain tissue in one embodiment;

[0059] Figure 7 A schematic diagram of the brain tissue surface area after the optimization of the veins and large blood vessels in one embodiment;

[0060] Figure 8is a schematic diagram of an internal segmentation result of an internal region of a brain tissue in one embodiment;

[0061] Fig. 9 is a schematic diagram of a sagittal plane of a target segmentation result of brain tissue in one embodiment;

[0062] Fig.10 is a schematic diagram of a coronal plane of a target segmentation result of brain tissue in one embodiment;

[0063] Fig.11 is a flowchart of a medical image segmentation method in another embodiment;

[0064] Fig.12 is a structural block diagram of a medical image segmentation device in one embodiment;

[0065] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0067] In one embodiment, Figure 1 As shown, a medical image segmentation method is provided. This embodiment uses the method applied to a terminal as an example. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server.

[0068] In this embodiment, the method includes the following steps:

[0069] Step S110, performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image.

[0070] The target part may be a part that needs to be segmented in a medical image, such as brain tissue.

[0071] The original image may be an originally acquired medical image, for example, a CTA image obtained by scanning brain tissue.

[0072] The mask image may be an image formed by a mask of the target part in the original image. For example, brain tissue is extracted from a CTA image, and the grayscale values ​​of the extracted pixels are set to 1 and the grayscale values ​​of other pixels are set to 0 to obtain a mask image of the brain tissue.

[0073] In a specific implementation, the terminal can segment the original image of the target part to obtain a mask image of the target part, perform morphological dilation processing on the mask image to obtain a dilated image of the target part, and can also perform morphological erosion processing on the mask image of the target part to obtain an eroded image of the target part.

[0074] In practical applications, the CTA image obtained by scanning the brain tissue can be used as the original image, denoted as image, loaded into the terminal, and the brain tissue in image can be segmented. The grayscale values ​​of the pixels in the segmented brain tissue area are set to 1, and the grayscale values ​​of other pixels are set to 0 to obtain a mask image of the brain tissue area, denoted as brain-mask. The brain-mask is expanded to obtain an expanded image mask1, and the brain-mask can also be corroded to obtain an eroded image mask2.

[0075] Step S120, determining the surface area of ​​the target part from the original image according to the dilation image and the erosion image, and determining the internal area of ​​the target part from the original image according to the erosion image.

[0076] The surface area may be a sulcus area outside the cerebral cortex.

[0077] The internal area may include the cerebral cortex, subcortical structures, and brainstem tissue.

[0078] In a specific implementation, the terminal can intercept the original image according to the intersection between the expansion image and the erosion image to obtain the surface area of ​​the target part, and can also intercept the original image according to the erosion image to obtain the internal area of ​​the target part.

[0079] It should be noted that the brain tissue includes the cerebral cortex composed of gray matter, white matter, and cerebrospinal fluid, subcortical structures, and brainstem tissue, and the external area of ​​the brain tissue includes brain tissue dura mater, external blood vessels, eyes, fat, muscles, and skull. Since the blood vessels in the sulcus area outside the cerebral cortex are easily missed during imaging, or are difficult to distinguish because they are too close to the skull, this application treats the sulcus area outside the cerebral cortex as the surface area and uses a segmentation method specifically for the surface area to segment it in order to improve the segmentation accuracy of the surface area, and treats the cerebral cortex composed of gray matter, white matter, cerebrospinal fluid, the subcortical structure, and the area where the brainstem tissue is located as the internal area, and uses a segmentation method specifically for the internal area to segment it in order to improve the segmentation accuracy of the internal area.

[0080] In practical applications, the intersection of mask1 and mask2 can be calculated to obtain mask3, and mask3 can be used to cut out the surface area of ​​the brain tissue from the image, recorded as zone1. Mask2 can also be directly used to cut out the internal area of ​​the brain tissue from the image, recorded as zone2.

[0081] Step S130, determining a surface segmentation result of the blood vessel in the surface area of ​​the target part, and determining an internal segmentation result of the blood vessel in the internal area.

[0082] In a specific implementation, the terminal can use different segmentation methods according to the characteristics of blood vessels in the surface area and internal area of ​​the target part, and perform blood vessel segmentation on the surface area and internal area respectively, so as to obtain the surface segmentation results of the blood vessels in the surface area of ​​the target part and the internal segmentation results in the internal area.

[0083] In practical applications, the blood vessels in zone1 can be segmented by, but are not limited to, a Gaussian mixture model, an adaptive threshold segmentation algorithm, a statistical model threshold segmentation algorithm, a graph cut algorithm, and an energy-based segmentation algorithm to obtain a blood vessel segmentation result vessel1 in the surface area of ​​the brain tissue, and a neural network, a deep learning (for example, a 3D-Unet network), and other methods are used to roughly segment the blood vessels in zone2. After optimizing the rough segmentation results (for example, filtering isolated noise points, smoothing edge areas, etc.), the blood vessel segmentation result vessel2 in the internal area of ​​the brain tissue is obtained.

[0084] Step S140, obtaining a target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result.

[0085] The target segmentation result may be a final segmentation result of the blood vessels at the target site.

[0086] In a specific implementation, the surface segmentation result and the internal segmentation result may be combined to obtain a target segmentation result of the blood vessels in the target area.

[0087] In practical applications, the blood vessel segmentation result vessel1 of the brain tissue surface area and the blood vessel segmentation result vessel2 of the brain tissue internal area can be combined to obtain the final blood vessel segmentation result vessel of the brain tissue area.

[0088] Figure 2 A flowchart of the blood vessel segmentation process in a brain tissue region is provided. Figure 2, the terminal can load a cerebral vascular image, denoted as image; use a brain tissue segmentation algorithm to segment the brain tissue in image to obtain a mask of the brain tissue area, denoted as brain-mask; then the brain-mask area can be expanded to obtain an expanded brain tissue mask mask1, and the brain-mask can be corroded to obtain an corroded brain tissue mask mask2; calculate the intersection of mask1 and mask2 to obtain a mask mask3 of the brain tissue surface area, and according to the intersection of mask3 and image, obtain the brain tissue surface area, denoted as zone1; calculate the intersection of mask2 and image to obtain the brain tissue internal area, denoted as zone2; perform vascular segmentation on zone1 to obtain vessel vessel1 in the brain tissue surface area; perform vascular segmentation on zone2 to obtain vessel vessel2 in the brain tissue internal area; merge vessel vessel1 and vessel vessel2 to obtain vessel vessel in the brain tissue area.

[0089] Figure 3 A flowchart of the blood vessel segmentation steps in the brain tissue internal area is provided. Figure 3 , it is possible but not limited to use 3D-Unet (three-dimensional U-shaped neural network) to segment the internal area zone2 of the brain tissue to obtain a rough segmentation result of the blood vessels in the internal area vessel21, and then vessel21 can be optimized, for example, isolated noise point filtering, edge smoothing and other optimization processes to obtain the blood vessel vessel2 in the internal area of ​​the brain tissue.

[0090] The above-mentioned medical image segmentation method obtains a morphological image of the target part by performing morphological processing on a mask image corresponding to an original image of the target part, wherein the morphological image includes a dilation image and an erosion image, and the surface area of ​​the target part is determined from the original image according to the dilation image and the erosion image, and the internal area of ​​the target part is determined from the original image according to the erosion image, and the surface segmentation result of the blood vessels in the surface area of ​​the target part is determined, and the internal segmentation result of the blood vessels in the internal area is determined, and the target segmentation result of the blood vessels is obtained according to the surface segmentation result and the internal segmentation result; the target part can be divided into a surface area and an internal area in the medical image, and different image segmentation processing is performed respectively according to the characteristics of the surface area and the internal area to improve the segmentation accuracy of each, and then by merging the surface segmentation result and the internal segmentation result, the segmentation accuracy of the entire target part is improved, and the accurate extraction of the target part is achieved.

[0091] In one embodiment, the above step S130 may specifically include: performing segmentation processing on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area; removing the first target from the first segmentation result to obtain a second segmentation result of the surface area; determining the second target from the second segmentation result, performing segmentation processing on the second target to obtain a third segmentation result of the surface area; merging the second segmentation result and the third segmentation result to obtain a surface segmentation result.

[0092] The segmentation model may be, but is not limited to, a Gaussian mixture model, an adaptive threshold segmentation model, a statistical threshold segmentation model, a graph cut model, an energy-based segmentation model, and the like.

[0093] The first segmentation result may be a segmentation result of a segmentation model, which includes bone residues, isolated blood vessel points, large veins on the surface of brain tissue, and small veins in grooves.

[0094] The first target may be, but is not limited to, bone residues and isolated blood vessel points near the skull. The second segmentation result may be a segmentation result after removing the bone residues and isolated blood vessel points near the skull.

[0095] The second target may be, but is not limited to, a vein or large blood vessel on the surface of the brain tissue. The third segmentation result may be a segmentation result of the vein or large blood vessel on the surface of the brain tissue.

[0096] In a specific implementation, a segmentation model can be pre-set on the terminal, and the segmentation model can be used to segment the surface area of ​​the target part to obtain a first segmentation result. The first targets such as bone residues and isolated blood vessel points can be removed from the first segmentation result to obtain a second segmentation result of the surface area. After that, the second targets such as veins and large blood vessels can be determined in the second segmentation result, and the second targets can be segmented to obtain a third segmentation result for the second targets in the surface area. The second segmentation result is merged with the third segmentation result to obtain a surface segmentation result of the target part.

[0097] Figure 4 A flowchart of the blood vessel segmentation steps of the brain tissue surface area is provided. Figure 4, the blood vessels of the brain tissue surface area zone1 can be enhanced to obtain the enhanced surface area image2; the blood vessels of image2 can be segmented by using the Gaussian mixture model segmentation algorithm to obtain the first segmentation result vessel11 of the brain tissue surface area; the bone residues near the skull and the isolated blood vessel points are removed from vessel11 to obtain the second segmentation result vessel12 of the brain tissue surface area; the large veins in vessel12 are obtained, and the large veins are optimized to obtain the third segmentation result vessel13 of the brain tissue surface area, wherein the optimization processing can be but is not limited to the optimization processing performed on the under-segmentation of the large veins; vessel12 is merged with vessel13 to obtain the fourth segmentation result vessel14 of the brain tissue surface area; the blood vessel contour and noise of vessel14 are optimized, including but not limited to boundary smoothing and noise point elimination, to obtain the blood vessel vessel1 of the brain tissue surface area.

[0098] In this embodiment, the surface area is segmented according to a preset segmentation model to obtain a first segmentation result of the surface area; the first target is removed from the first segmentation result to obtain a second segmentation result of the surface area; the second target is determined from the second segmentation result, and the second target is segmented to obtain a third segmentation result of the surface area; the second segmentation result and the third segmentation result are merged to obtain a surface segmentation result, which can eliminate bone residues and isolated blood vessel points near the skull from the segmentation result of the surface area of ​​the target part, solve the problem of under-segmentation of large blood vessels near the skull, and improve the accuracy of blood vessel segmentation in the surface area of ​​the target part.

[0099] In one embodiment, the above-mentioned step of segmenting the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area may specifically include: enhancing the blood vessels in the surface area to obtain an enhanced surface area; and segmenting the enhanced surface area according to the segmentation model to obtain the first segmentation result.

[0100] In a specific implementation, the terminal can perform vascular enhancement on the surface area of ​​the target part to remove bone fragments near the skull outside the brain tissue to obtain an enhanced surface area, and can also use a Gaussian mixture model to perform vascular segmentation on the enhanced surface area to obtain a first segmentation result of the blood vessels in the target part.

[0101] In practical applications, in order to remove bone fragments near the skull outside the brain tissue, the blood vessels in the brain tissue surface area zone1 can be enhanced to obtain the enhanced image image2. The Gaussian mixture model segmentation algorithm can also be used to perform blood vessel segmentation on the enhanced image image2 to obtain the blood vessel vessel11 in the brain tissue surface area.

[0102] In this embodiment, the blood vessels in the surface area are enhanced to obtain an enhanced surface area; the enhanced surface area is segmented according to the segmentation model to obtain a first segmentation result, and the blood vessels in the surface area of ​​the target part can be enhanced and preliminarily segmented to improve the segmentation efficiency of the surface area.

[0103] In one embodiment, the above step S130 may further specifically include: determining a connected domain in the first segmentation result; when a property of the connected domain meets a first preset condition, determining the connected domain as the first target; the property includes at least one of shape and volume, and the first preset condition includes at least one of shape abnormality and volume abnormality.

[0104] In a specific implementation, the terminal may identify connected domains in the first segmentation result, and determine the shape and / or volume of each connected domain. If a connected domain has an abnormal shape and / or abnormal volume, the connected domain is determined as the first target.

[0105] In practical applications, the use of a Gaussian mixture model to segment blood vessels in the brain tissue surface area is prone to certain over-segmentation phenomena, so vessel11 needs to be optimized, including removing bone residues near the skull and isolated blood vessel points from vessel11. Specifically, a connected domain analysis can be performed on vessel11, and the connected blood vessels in vessel11 can be divided into multiple separate connected domains. The second-order moment and volume size of each connected domain are calculated, wherein the shape of the connected domain can be judged whether it is abnormal based on the second-order moment of the connected domain, and a second-order moment threshold and a volume threshold are set. The connected domains with abnormal shapes are filtered out based on the second-order moment threshold, and the connected domains with smaller volumes are filtered out based on the volume threshold, thereby obtaining the optimized brain tissue surface area blood vessel12.

[0106] In this embodiment, by determining the connected domain in the first segmentation result; when the properties of the connected domain meet the first preset condition, the connected domain is determined as the first target, bone residues near the skull and isolated blood vessel points can be removed from the first segmentation result, thereby improving the accuracy of the brain tissue surface segmentation result, and further improving the accuracy of brain tissue segmentation.

[0107] In one embodiment, the above-mentioned step of determining the second target from the second segmentation result, segmenting the second target, and obtaining the third segmentation result of the surface area may specifically include: determining the maximum connected domain in the second segmentation result; dilating the maximum connected domain to obtain the second target; and segmenting the blood vessels in the second target to obtain the third segmentation result.

[0108] In a specific implementation, the terminal can identify the largest connected domain in the second segmentation result, dilate the largest connected domain, determine the expanded largest connected domain as the second target, perform blood vessel segmentation on the second target, and obtain a third segmentation result of the surface area of ​​the target part.

[0109] In practical applications, since there is a certain under-segmentation problem with the large blood vessels near the skull, the large veins in vessel12 can be further segmented to solve the under-segmentation problem. Specifically, the terminal can obtain the maximum connected domain connect in vessel12. Since the veins are the largest blood vessels on the surface of brain tissue, the maximum connected domain in vessel12 corresponds to the veins (for example, the superior sagittal sinus, sigmoid sinus and transverse sinus). Then, the maximum connected domain connect can be expanded to obtain the second target connect1, and the intersection of connect1 and image is calculated to obtain image3; the adaptive threshold segmentation algorithm is used to perform vascular segmentation on image3 to obtain the optimized venous vessel vessel13, which is the third segmentation result.

[0110] In this embodiment, by determining the maximum connected domain in the second segmentation result; expanding the maximum connected domain to obtain the second target; and segmenting the blood vessels in the second target to obtain the third segmentation result, the problem of under-segmentation of large blood vessels near the skull can be solved, and the accuracy of the brain tissue surface segmentation results can be further improved, thereby improving the accuracy of brain tissue segmentation.

[0111] In one embodiment, the above step S120 may specifically include: obtaining a surface mask image of the target part according to the intersection of the dilation image and the erosion image; and performing mask processing on the original image according to the surface mask image to obtain a surface area of ​​the target part.

[0112] In a specific implementation, the terminal may perform an intersection operation on the expansion image and the erosion image to obtain a surface mask image of the target part, and perform an intersection operation on the surface mask image and the original image to obtain a surface area of ​​the target part.

[0113] In practical applications, the intersection of the expansion image mask1 and the erosion image mask2 can be calculated to obtain the surface mask image mask3, and the intersection of the surface mask image mask3 and the original image image can also be calculated to obtain the brain tissue surface area zone1.

[0114] In this embodiment, a surface mask image of the target part is obtained according to the intersection of the dilation image and the erosion image; the original image is masked according to the surface mask image to obtain the surface area of ​​the target part. The surface area of ​​the target part can be quickly extracted from the original image, thereby improving the efficiency of image segmentation.

[0115] In one embodiment, the above step S120 may further specifically include: obtaining an internal mask image of the target part according to the corrosion image; and performing mask processing on the original image according to the internal mask image to obtain an internal area of ​​the target part.

[0116] In a specific implementation, the terminal may determine the erosion image as the internal mask image of the target part, perform an intersection operation on the internal mask image and the original image, and obtain the internal area of ​​the target part.

[0117] In practical applications, the intersection of the eroded image mask2 and the original image image can be calculated to obtain the internal area zone2 of the brain tissue.

[0118] In this embodiment, an internal mask image of the target part is obtained based on the corrosion image; the original image is masked according to the internal mask image to obtain the internal area of ​​the target part. The internal area of ​​the target part can be quickly extracted from the original image, thereby improving the efficiency of image segmentation.

[0119] In order to facilitate those skilled in the art to have a deeper understanding of the embodiments of the present application, a specific example will be described below.

[0120] In order to accurately extract the blood vessels inside and on the surface of brain tissue in CTA images, this application proposes a CTA intracranial blood vessel segmentation method, which specifically includes the following steps:

[0121] Step 100, loading a cerebral vascular image, ie, an original image, denoted as image.

[0122] Step 200, using a brain tissue segmentation algorithm to segment the brain tissue in the image, and obtaining a mask image of the brain tissue area, recorded as brain-mask.

[0123] Step 300: dilate the brain-mask region to obtain a dilated image mask1, and erode the brain-mask region to obtain an erode image mask2.

[0124] Step 400, calculating the intersection of mask1 and mask2 to obtain a surface mask image mask3.

[0125] Step 500, calculate the intersection of mask3 and image, and get Figure 5 The brain tissue surface area shown is recorded as zone1; calculate the intersection of mask2 and image, and get Figure 6 The internal area of ​​the brain tissue shown is marked as zone2.

[0126] Step 600, performing blood vessel segmentation on zone 1 to obtain a surface segmentation result vessel 1 of the brain tissue surface area.

[0127] Specifically, the blood vessel segmentation algorithm of the brain tissue surface area may include the following steps:

[0128] Step 610 , performing blood vessel enhancement on the brain tissue surface area zone1 to obtain an enhanced surface area image2 .

[0129] The purpose of vascular augmentation is to remove bone fragments near the skull outside the brain tissue.

[0130] Step 620 , use a Gaussian mixture model segmentation algorithm to perform blood vessel segmentation on image2 to obtain a first segmentation result vessel11 of the brain tissue surface area.

[0131] Among them, the Gaussian mixture model is a vascular Gaussian model that combines two Gaussian distributions to segment the blood vessels in the surface area of ​​​​the brain tissue. This method has a certain degree of over-segmentation, so vessel11 needs to be optimized later.

[0132] Step 630 , analyzing vessel 11 , removing bone residues near the skull and isolated blood vessel points, and obtaining a second segmentation result vessel 12 of the brain tissue surface area.

[0133] Specifically, a connected domain analysis can be performed on vessel11, and the connected blood vessels therein can be divided into multiple separate connected domains. The second-order moment (used to determine the shape) and volume of each connected domain can be calculated, and a threshold can be set to filter out connected domains with small volumes and abnormal shapes to obtain a second segmentation result vessel12.

[0134] Step 640, obtain the veins and large blood vessels in vessel 12, optimize the veins and large blood vessels to obtain the third segmentation result vessel 13 of the brain tissue surface area, such as Figure 7 shown.

[0135] Specifically, the blood vessels near the brain tissue are close to the skull, which may cause some under-segmentation of the large blood vessels near the skull. The following methods can be used to deal with the under-segmentation problem:

[0136] Step 641, obtaining the maximum connected domain connect in vessel12. The vein is the largest blood vessel on the surface of the brain tissue, so the maximum connected domain in vessel12 corresponds to the vein (superior sagittal sinus, sigmoid sinus and transverse sinus);

[0137] Step 642, expand the maximum connected domain connect to obtain connect1, calculate the intersection of connect1 and image, and obtain image3;

[0138] Step 643, using an adaptive threshold segmentation algorithm, perform blood vessel segmentation on image3 to obtain a third segmentation result vessel13 of the brain tissue surface area.

[0139] Step 650, merging vessel 12 and vessel 13 to obtain a fourth segmentation result vessel 14 of the brain tissue surface area;

[0140] Step 660 , optimizing the contour of vessel 14 to obtain a surface segmentation result vessel 1 of the brain tissue surface area.

[0141] The optimization process includes but is not limited to boundary smoothing and noise point removal.

[0142] Step 700, performing blood vessel segmentation on zone 2 to obtain an internal segmentation result vessel2 of the internal area of ​​the brain tissue.

[0143] Specifically, the blood vessel segmentation algorithm for the internal area of ​​the brain tissue may include the following steps:

[0144] Step 710 , using a 3D-Unet network to segment the blood vessels in the brain tissue internal region zone2 to obtain a roughly segmented blood vessel vessel21 .

[0145] Among them, the 3D-Unet network belongs to a deep learning method, which easily leads to under-segmentation problems in the edges of the segmented images, so it is necessary to further optimize vessel 21. It should be noted that the segmentation method here is not limited to the 3D-Unet network.

[0146] Step 720, optimize vessel21 to obtain the internal segmentation result vessel2 of the internal area of ​​the brain tissue, such as Figure 8 shown.

[0147] The optimization method includes but is not limited to filtering isolated noise points in the vessel 21 and smoothing the edge area of ​​the vessel 21 .

[0148] Step 800, merging vessel1 and vessel2 to obtain the target segmentation result vessel of the blood vessels in the brain tissue area, such as Fig. 9 and Fig.10 shown.

[0149] The above-mentioned CTA intracranial blood vessel segmentation method can, on the one hand, more accurately extract the blood vessels inside the brain tissue, providing a reference for the formulation of surgical plans and disease diagnosis. On the other hand, the method searches for the transition zone of the blood vessel edge inside the brain tissue, improves the problem of under-segmentation, and removes the bone pieces outside the brain tissue, and repairs the large veins on the surface of the brain tissue, thereby improving the accuracy of CTA brain tissue surface blood vessel segmentation. Specifically, by dividing the brain blood vessels into the internal area of ​​the brain tissue and the surface area of ​​the brain tissue, and processing each area separately, the bone residue fragments near the surface of the brain tissue can be effectively removed. Moreover, by optimizing the vein area near the surface of the brain tissue, the problem of under-segmentation of the veins near the skull is effectively improved. In addition, by searching for the transition zone around the blood vessels inside the brain tissue, the problem of under-segmentation of the blood vessels inside the brain tissue in the transition zone is further optimized. Compared with the prior art, the above-mentioned CTA intracranial blood vessel segmentation method has a certain adaptability. While segmenting the CTA blood vessel image, it can not only ensure the richness of the blood vessels, but also eliminate the interference of the skull on the surface of the brain tissue as much as possible.

[0150] In one embodiment, Fig.11 As shown, a medical image segmentation method is provided, which is described by taking the method applied to a terminal as an example, and includes the following steps:

[0151] Step S201, performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0152] Step S202, determining the surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining the internal area of ​​the target part from the original image according to the eroded image;

[0153] Step S203, performing segmentation processing on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area;

[0154] Step S204, removing the first target from the first segmentation result to obtain a second segmentation result of the surface area;

[0155] Step S205, determining a second target from the second segmentation result, performing segmentation processing on the second target, and obtaining a third segmentation result of the surface area;

[0156] Step S206, merging the second segmentation result and the third segmentation result to obtain a surface segmentation result;

[0157] Step S207, determining the internal segmentation result of the blood vessel in the internal area;

[0158] Step S208, obtaining a target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result.

[0159] In a specific implementation, the terminal can segment the original image of the target part to obtain a mask image. The mask image is morphologically processed to obtain a dilated image and an eroded image of the target part. According to the intersection of the dilated image and the eroded image, the surface area of ​​the target part is extracted from the original image. According to the eroded image, the internal area of ​​the target part is extracted from the original image. Then, the segmentation model can be used to segment the surface area to obtain a first segmentation result. The first target in the first segmentation result is eliminated to obtain a second segmentation result. The second target in the second segmentation result is further segmented to obtain a third segmentation result. The second segmentation result is merged with the third segmentation result to obtain a surface segmentation result of the blood vessels in the target part. Then, the internal area of ​​the target part is segmented to obtain an internal segmentation result of the blood vessels. The surface segmentation result is merged with the internal segmentation result to obtain a target segmentation result of the blood vessels in the target part.

[0160] In this embodiment, a morphological image of the target part is obtained by performing morphological processing on a mask image corresponding to an original image of the target part; the morphological image includes a dilation image and an erosion image; a surface area of ​​the target part is determined from the original image according to the dilation image and the erosion image; and an internal area of ​​the target part is determined from the original image according to the erosion image; a segmentation process is performed on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area; a first target is removed from the first segmentation result to obtain a second segmentation result of the surface area; a second target is determined from the second segmentation result; a segmentation process is performed on the second target to obtain a third segmentation result of the surface area; the second segmentation result and the third segmentation result are merged to obtain a surface segmentation result; an internal segmentation result of a blood vessel in the internal area is determined; and a target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result; the target part can be divided into a surface area and an internal area in a medical image, and different image segmentation processes are performed respectively according to the characteristics of the surface area and the internal area to improve their respective segmentation accuracies; and then the segmentation accuracy of the entire target part is improved by merging the surface segmentation result and the internal segmentation result, so as to achieve accurate extraction of the target part.

[0161] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0162] Based on the same inventive concept, the embodiment of the present application also provides a medical image segmentation device for implementing the medical image segmentation method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in one or more medical image segmentation device embodiments provided below can refer to the limitations of the medical image segmentation method above, and will not be repeated here.

[0163] In one embodiment, Fig.12 As shown, a medical image segmentation device is provided, including: a processing module 310, a determination module 320, a segmentation module 330 and a merging module 340, wherein:

[0164] The processing module 310 is used to perform morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image;

[0165] A determination module 320, configured to determine a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and to determine an internal area of ​​the target part from the original image according to the eroded image;

[0166] A segmentation module 330, configured to determine a surface segmentation result of the blood vessel at the target site in the surface region, and to determine an internal segmentation result of the blood vessel in the internal region;

[0167] The merging module 340 is used to obtain the target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result.

[0168] In one embodiment, the above-mentioned segmentation module 330 is also used to perform segmentation processing on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area; remove the first target from the first segmentation result to obtain a second segmentation result of the surface area; determine the second target from the second segmentation result, perform segmentation processing on the second target to obtain a third segmentation result of the surface area; merge the second segmentation result and the third segmentation result to obtain the surface segmentation result.

[0169] In one embodiment, the segmentation module 330 is further used to enhance the blood vessels in the surface area to obtain an enhanced surface area; and segment the enhanced surface area according to the segmentation model to obtain the first segmentation result.

[0170] In one embodiment, the above-mentioned segmentation module 330 is also used to determine a connected domain in the first segmentation result; when the property of the connected domain meets the first preset condition, the connected domain is determined as the first target; the property includes at least one of shape and volume, and the first preset condition includes at least one of shape abnormality and volume abnormality.

[0171] In one embodiment, the segmentation module 330 is further used to determine the maximum connected domain in the second segmentation result; dilate the maximum connected domain to obtain the second target; and segment the blood vessels in the second target to obtain the third segmentation result.

[0172] In one embodiment, the determination module 320 is further used to obtain a surface mask image of the target part according to the intersection of the dilation image and the erosion image; and to perform mask processing on the original image according to the surface mask image to obtain the surface area of ​​the target part.

[0173] In one embodiment, the determination module 320 is further used to obtain an internal mask image of the target part according to the erosion image; and perform mask processing on the original image according to the internal mask image to obtain the internal area of ​​the target part.

[0174] Each module in the above-mentioned medical image segmentation device can be implemented in whole or in part by software, hardware or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0175] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.13 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a medical image segmentation method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0176] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0177] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0178] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0179] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0180] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0181] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0182] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0183] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A medical image segmentation method, characterized in that: The method comprises: Performing morphological processing on the mask image corresponding to the original image of the target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image; Determining a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and determining an internal area of ​​the target part from the original image according to the eroded image; Determine a surface segmentation result of the blood vessel at the target site in the surface area, and determine an internal segmentation result of the blood vessel in the internal area; A target segmentation result of the blood vessel is obtained according to the surface segmentation result and the internal segmentation result.

2. The method according to claim 1, characterized in that Determining the surface segmentation result of the blood vessel at the target site in the surface area includes: Performing segmentation processing on the surface area according to a preset segmentation model to obtain a first segmentation result of the surface area; removing the first target from the first segmentation result to obtain a second segmentation result of the surface area; determining a second target from the second segmentation result, performing segmentation processing on the second target, and obtaining a third segmentation result of the surface area; The second segmentation result and the third segmentation result are combined to obtain the surface segmentation result.

3. The method according to claim 2, characterized in that The step of performing segmentation processing on the surface area according to the preset segmentation model to obtain a first segmentation result of the surface area includes: Performing enhancement processing on the blood vessels in the surface area to obtain an enhanced surface area; The enhanced surface area is segmented according to the segmentation model to obtain the first segmentation result.

4. The method according to claim 2, characterized in that: Before removing the first target from the first segmentation result to obtain a second segmentation result of the surface area, the method further includes: Determining a connected domain in the first segmentation result; When the property of the connected domain meets a first preset condition, the connected domain is determined as the first target; the property includes at least one of shape and volume, and the first preset condition includes at least one of abnormal shape and abnormal volume.

5. The method according to claim 2, characterized in that: The determining a second target from the second segmentation result, performing segmentation processing on the second target, and obtaining a third segmentation result of the surface area includes: Determining a maximum connected domain in the second segmentation result; Performing expansion processing on the maximum connected domain to obtain the second target; The blood vessel in the second target is segmented to obtain the third segmentation result.

6. The method according to claim 1, characterized in that Determining the surface area of ​​the target part from the original image according to the dilated image and the eroded image includes: Obtaining a surface mask image of the target part according to the intersection of the dilation image and the erosion image; The original image is subjected to mask processing according to the surface mask image to obtain the surface area of ​​the target part.

7. The method according to claim 1, characterized in that Determining the inner area of ​​the target part from the original image according to the eroded image includes: Obtaining an internal mask image of the target part according to the corrosion image; The original image is subjected to mask processing according to the internal mask image to obtain the internal area of ​​the target part.

8. A medical image segmentation device, characterized in that: The device comprises: A processing module, used for performing morphological processing on a mask image corresponding to an original image of a target part to obtain a morphological image of the target part; the morphological image includes a dilation image and an erosion image; a determination module, configured to determine a surface area of ​​the target part from the original image according to the dilated image and the eroded image, and to determine an internal area of ​​the target part from the original image according to the eroded image; a segmentation module, used to determine a surface segmentation result of the blood vessel at the target site in the surface area, and to determine an internal segmentation result of the blood vessel in the internal area; A merging module is used to obtain a target segmentation result of the blood vessel according to the surface segmentation result and the internal segmentation result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images