An image processing method based on axillary blood vessel-lymph node DCE-MRI images

Abstract

This application relates to an image processing method based on axillary vascular-lymph node DCE-MRI images, comprising: preprocessing an initial axillary vascular-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary vascular-lymph node image; constructing a first maximum likelihood function based on the Hessian matrix, and extracting vascular target images of a preset time series based on the first maximum likelihood function; constructing a second maximum likelihood function based on the Hessian matrix, determining the location of lymph nodes in the initial axillary vascular-lymph node image based on the second maximum likelihood function, and extracting lymph node target images of a preset time series using a three-dimensional region growing algorithm; and plotting a vascular-lymph node TIC curve by calculating the gray-level mean of the images based on the preset time series vascular target images and lymph node target images. In this application, the difference in the area of ​​the obtained TIC curves is used to determine whether metastasis has occurred in the axillary lymph nodes of breast cancer.

Description

Technical Field

[0001] The present application belongs to the field of image processing technology, and specifically relates to an image processing method based on axillary blood vessel-lymph node DCE-MRI images. Background Art

[0002] Axillary lymph nodes are the earliest and most common metastatic site of breast cancer. Axillary lymph node metastasis and its load status are of great value in the selection of surgical methods for breast cancer, the formulation of adjuvant treatment plans, and the evaluation of post-healing. Imaging examination, as a non-invasive examination method with no or little harm, is the most common and necessary diagnostic link for breast diseases in clinical applications. Therefore, using DCE-MRI images to propose a method for determining axillary lymph node metastasis of breast cancer is of great clinical value.

[0003] At present, the diagnosis of breast cancer is mainly given by doctors in the imaging department based on their professional knowledge and rich clinical experience. The sensitivity, specificity and accuracy need to be improved. At the same time, it takes a lot of doctors' time and is inefficient. Currently, sentinel lymph node biopsy and axillary lymph node dissection are used in clinical practice to evaluate the metastasis status of axillary lymph nodes and clarify the stage. However, this invasive method can lead to multiple complications such as lymphedema, pain, numbness, and movement disorders, causing unnecessary pain to patients who have not yet developed a high lymph node metastasis load.

[0004] In view of this, the present application provides an image processing method based on axillary blood vessel-lymph node DCE-MRI images, which can achieve accurate three-dimensional segmentation of axillary blood vessels and lymph nodes, and then judge the axillary lymph node metastasis of breast cancer. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides an image processing method based on axillary blood vessel-lymph node DCE-MRI images.

[0007] (II) Technical solution

[0008] In order to achieve the above objectives, this application adopts the following technical solutions:

[0009] In a first aspect, the present application provides an image processing method based on axillary blood vessel-lymph node DCE-MRI images, comprising:

[0010] S1, preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image;

[0011] The preprocessing includes: adjusting the window level for the initial axillary blood vessel-lymph node image;

[0012] The initial axillary blood vessel-lymph node image is an image processed with a contrast agent, including a blood vessel target image and a lymph node target image;

[0013] S2. constructing a first maximum likelihood function based on the Hessian matrix, and extracting a blood vessel target image of a preset time series based on the first maximum likelihood function;

[0014] S3, constructing a second maximum likelihood function based on the Hessian matrix, determining the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extracting the lymph node target image of a preset time series by a three-dimensional region growing algorithm;

[0015] S4, based on the vascular target image and the lymph node target image of the preset time series, drawing a vascular-lymph node TIC curve diagram by calculating the grayscale mean value corresponding to the image;

[0016] The blood vessel-lymph node TIC curve diagram includes a blood vessel TIC curve and a lymph node TIC curve.

[0017] Optionally, also include:

[0018] S5. Determine the metastasis of axillary lymph nodes of breast cancer by obtaining the mean, standard deviation and median of the TIC curve difference;

[0019] The TIC curve difference is: the difference between the area under the vascular TIC curve and the area under the lymph node TIC curve in the vascular-lymph node TIC curve diagram.

[0020] Optionally, the step of obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image in S1 includes:

[0021] Obtain eigenvalues ​​λ1, λ2, and λ3 of the Hessian matrix, and sort the obtained eigenvalues ​​based on a preset eigenvalue sorting condition;

[0022] The preset characteristic value sorting condition is: |λ1|≤|λ2|≤|λ3|;

[0023] The Hessian matrix is:

[0024]

[0025] Among them, x, y, z represent the three directions of x, y, and z, and f means finding the derivative.

[0026] Optionally, constructing a first maximum likelihood function based on the Hessian matrix in S2 includes:

[0027] The first maximum likelihood function is:

[0028]

[0029] Among them, α1, β1, and γ1 are parameters that control the sensitivity of the calculation process of the first maximum likelihood function;

[0030] in,

[0031] Optionally, constructing a second maximum likelihood function based on the Hessian matrix in S3 includes:

[0032] The second maximum likelihood function is:

[0033]

[0034] Among them, α2, β2, and γ2 are parameters that control the sensitivity of the second maximum likelihood function calculation process.

[0035] Optionally, the determining the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function in S3 includes:

[0036] Based on the initial axillary blood vessel-lymph node image, the layer where the lymph nodes are located is preliminarily extracted through the second maximum likelihood function;

[0037] Taking the initially extracted layer as the central layer, one layer is obtained above and below the central layer, and based on the central layer and the layers above and below the central layer, the layer with the brightest grayscale value is obtained; and the location of the lymph node is determined based on the layer with the brightest grayscale value.

[0038] Optionally, the step of extracting lymph node target images of a preset time series by using a three-dimensional region growing algorithm in S3 includes:

[0039] S3-1, calculating the absolute value of the corresponding pixel difference between the initial pixel and the surrounding adjacent pixels, and establishing a model in which all pixel values ​​are 0;

[0040] S3-2, according to the preset pixel difference threshold and the absolute value of the pixel difference in step S3-1, the corresponding pixel value in the established model is set to 1;

[0041] S3-3. For the pixel points whose pixel values ​​are set to 1 in the model, repeat the above steps S3-1 and S3-2 until the absolute values ​​of the pixel point differences between all the pixel points and the surrounding adjacent pixel points are calculated.

[0042] Optionally, the step of drawing a blood vessel-lymph node TIC curve diagram based on the blood vessel target image and the lymph node target image in a preset time series by calculating the grayscale mean corresponding to the image includes:

[0043] Based on the vascular target image and the lymph node target image obtained after contrast agent treatment, the grayscale mean values ​​corresponding to multiple preset time points are obtained for the area to be studied in the image;

[0044] The multiple preset time points are time points at the same time interval.

[0045] In a second aspect, the present application provides an image processing system based on axillary blood vessel-lymph node DCE-MRI images, comprising:

[0046] An image preprocessing module, specifically used for preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image;

[0047] A blood vessel image acquisition module, specifically used to construct a first maximum likelihood function based on the Hessian matrix, and extract a blood vessel target image of a preset time series based on the first maximum likelihood function;

[0048] A lymph node image acquisition module, specifically used to construct a second maximum likelihood function based on the Hessian matrix, determine the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extract lymph node target images of a preset time series through a three-dimensional region growing algorithm;

[0049] The TIC image acquisition module is specifically used to draw a blood vessel-lymph node TIC curve diagram by calculating the grayscale mean value corresponding to the image based on the blood vessel target image and the lymph node target image of the preset time series.

[0050] (III) Beneficial effects

[0051] This application achieves accurate three-dimensional segmentation of axillary blood vessels and lymph nodes by acquiring blood vessel target images and lymph node images, and draws a blood vessel-lymph node TIC curve based on the acquired blood vessel target images and lymph node images to judge and predict the metastasis of axillary lymph nodes of breast cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The present application is described with the aid of the following drawings:

[0053] Figure 1 is a flowchart of an image processing method based on axillary blood vessel-lymph node DCE-MRI images;

[0054] Figure 2-1This is the initial axillary vascular-lymph node image;

[0055] Figure 2-2 The initial axillary vascular-lymph node image after preprocessing;

[0056] Figure 3-1 is the vascular portion image corresponding to the initial axillary vascular-lymph node image;

[0057] Figure 3-2 is the vascular target image obtained after segmentation;

[0058] Figure 4-1 It is the partial lymph node image corresponding to the initial axillary blood vessel-lymph node image;

[0059] Figure 4-2 is the target image of lymph nodes obtained after segmentation;

[0060] Figure 5-1 The blood vessel-lymph node TIC curve corresponding to lymph node metastasis;

[0061] Figure 5-2 This is the TIC curve of blood vessels and lymph nodes corresponding to the absence of lymph node metastasis;

[0062] Figure 6 This is a diagram showing the module composition of an image processing system based on axillary blood vessel-lymph node DCE-MRI images;

[0063] Figure 7 A schematic diagram of the composition of an electronic device. DETAILED DESCRIPTION

[0064] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below in conjunction with the accompanying drawings through specific implementation methods. It is understood that the specific embodiments described below are only used to explain the relevant inventions, rather than to limit the invention. It should also be noted that the embodiments and features in the embodiments of this application can be combined with each other in the absence of conflict; for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0065] Before describing the image processing method and system based on axillary blood vessel-lymph node DCE-MRI images disclosed in this application through specific embodiments, it should be noted that:

[0066] The present application is a computer program, which can implement corresponding functions based on electronic devices (such as computers, laptops, etc.), and perform corresponding processing on the initial axillary blood vessel-lymph node image by receiving operation instructions. The processing process of the initial axillary blood vessel-lymph node image refers to the image processing method in the specific embodiment.

[0067] Embodiment 1 provides an image processing method based on axillary blood vessel-lymph node DCE-MRI images, such as Figure 1 As shown, the details are as follows:

[0068] S1. Preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image.

[0069] In this embodiment, the preprocessing includes: adjusting the window level for the initial axillary blood vessel-lymph node image.

[0070] In this embodiment, the initial axillary blood vessel-lymph node image is an image processed with a contrast agent, including a blood vessel target image and a lymph node target image.

[0071] S2. Constructing a first maximum likelihood function based on the Hessian matrix, and extracting a blood vessel target image of a preset time series based on the first maximum likelihood function.

[0072] S3. Construct a second maximum likelihood function based on the Hessian matrix, determine the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extract the lymph node target image of a preset time series through a three-dimensional region growing algorithm.

[0073] S4. Based on the vascular target image and lymph node target image of the preset time series, a vascular-lymph node TIC curve diagram is drawn by calculating the grayscale mean values ​​corresponding to the images.

[0074] In this embodiment, the blood vessel-lymph node TIC curve diagram includes a blood vessel TIC curve and a lymph node TIC curve.

[0075] The image processing method in Example 1 discloses that a Hessian matrix is ​​constructed based on an initial axillary blood vessel-lymph node image, a blood vessel target image is extracted based on the Hessian matrix, and a lymph node target image is extracted based on the Hessian matrix through a three-dimensional region growing algorithm, so as to achieve accurate three-dimensional segmentation of axillary blood vessels and lymph nodes, and a blood vessel-lymph node TIC curve is drawn based on the acquired blood vessel target image and lymph node image to judge and predict the metastasis of axillary lymph nodes of breast cancer.

[0076] Embodiment 2 provides an image processing method based on axillary blood vessel-lymph node DCE-MRI images, which is as follows:

[0077] S1. Preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image.

[0078] Based on the aforementioned step S1, for the preprocessing of the initial axillary blood vessel-lymph node image, it should be noted that:

[0079] For the initial axillary blood vessel-lymph node image in the form of DCE-MRI images, since different tissues in the image absorb the contrast agent at different rates during image scanning, the image grayscale values ​​are not the same and the grayscale value range is large, so the image window level needs to be adjusted to ensure that the correct position is found when the image is selected. Specifically, Figure 2-1 For the initial axillary vascular-lymph node image, Figure 2-2 This is the initial axillary vascular-lymph node image after preprocessing.

[0080] Based on the above step S1, for the obtained Hessian matrix, it should be noted that:

[0081] In this embodiment, for obtaining the aforementioned Hessian matrix, the Gaussian kernel is set to 1;

[0082] The Hessian matrix is ​​specifically expressed as:

[0083]

[0084] Among them, x, y, z represent the three directions of x, y, and z, and f means finding the derivative.

[0085] In this embodiment, a processing method based on the Hessian matrix includes: obtaining eigenvalues ​​λ1, λ2, and λ3 of the Hessian matrix, and sorting the obtained eigenvalues ​​based on a preset eigenvalue sorting condition.

[0086] In this embodiment, the preset characteristic value sorting condition is: |λ1|≤|λ2|≤|λ3|.

[0087] S2. Constructing a first maximum likelihood function based on the Hessian matrix, and extracting a blood vessel target image of a preset time series based on the first maximum likelihood function.

[0088] Based on the aforementioned step S2, it should be noted that the first maximum likelihood function is constructed based on the acquired eigenvalues ​​λ1, λ2, and λ3 of the Hessian matrix. The constructed first maximum likelihood function is specifically as follows:

[0089]

[0090] In the above formula, α1, β1, and γ1 are parameters for controlling the sensitivity of the first maximum likelihood function calculation process; in this embodiment, α1=0.5, β1=0.5, and γ1=15.

[0091] In this embodiment, based on the eigenvalues ​​λ1, λ2, and λ3 of the aforementioned Hessian matrix, the blood vessel region in the initial axillary blood vessel-lymph node image is analyzed. It should be noted that:

[0092] First, for an ideal blood vessel shape in a blood vessel region, the following conditions must be satisfied: |λ1|≈0, |λ1|<<|λ2|, |λ2|≈|λ3|.

[0093] Second, the plate-like and linear-like blood vessel shapes in the blood vessel area are distinguished by calculating the aforementioned R A Make a judgment, specifically: if |λ2|>|λ3| is satisfied, the shape of the blood vessel is plate-like, and if |λ2|≤|λ3| is satisfied, the shape of the blood vessel is linear.

[0094] Third, the spot-like blood vessel shapes in the blood vessel area are distinguished by calculating the aforementioned R B Make a judgment, specifically: if |λ1| and |λ2| approach infinitesimal at the same time, the corresponding area is approximately a point; if |λ2| and |λ3| approach infinitesimal at the same time, the corresponding area is a linear spot or a plate-like spot.

[0095] Fourth, since the grayscale value of the blood vessel area in the initial axillary blood vessel-lymph node image is obviously higher than the grayscale value of the background part, the background part and the blood vessel part are distinguished by calculating the aforementioned S.

[0096] In this embodiment, Figure 3-1 and Figure 3-2 As shown, the corresponding blood vessel target image is extracted using the aforementioned first maximum likelihood function, where: Figure 3-1 is the vascular part image corresponding to the initial axillary vascular-lymph node image, Figure 3-2 is the vascular target image obtained after segmentation.

[0097] S3. Construct a second maximum likelihood function based on the Hessian matrix, determine the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extract the lymph node target image of a preset time series through a three-dimensional region growing algorithm.

[0098] Based on the aforementioned step S3, it should be noted that the second maximum likelihood function is constructed based on the acquired eigenvalues ​​λ1, λ2, and λ3 of the Hessian matrix. The constructed second maximum likelihood function is specifically as follows:

[0099]

[0100] In the above formula, α2, β2, and γ2 are parameters for controlling the sensitivity of the second maximum likelihood function calculation process; in this embodiment, α2=0.5, β2=0.5, and γ2=15.

[0101] Based on the aforementioned step S3, for determining the location of the lymph nodes in the initial axillary blood vessel-lymph node image, it should be noted that:

[0102] First, based on the initial axillary blood vessel-lymph node image, the layer where the lymph nodes are located is preliminarily extracted through the second maximum likelihood function.

[0103] Second, a layer is obtained above and below the center layer using the initially extracted layer, and a layer with the brightest gray value is obtained based on the center layer and the layers above and below the center layer.

[0104] Third, determine the location of the lymph nodes based on the layer with the brightest grayscale value.

[0105] Based on the aforementioned step S3, for extracting the lymph node target image by the three-dimensional region growing algorithm, it should be noted that:

[0106] First, the corresponding pixel difference values ​​are calculated for the initial pixel and the surrounding adjacent pixels, and a model is established in which all pixel values ​​are 0.

[0107] Second, according to the preset pixel difference threshold and the pixel difference value in step S3-1, the corresponding pixel value in the established model is set to 1.

[0108] Third, for the pixel points in the model whose pixel values ​​are set to 1, repeat the above steps S3-1 and S3-2 until the pixel point differences between all the pixel points and the surrounding adjacent pixel points are calculated.

[0109] Based on the above three-dimensional region growing algorithm, it should be noted that: in the process of acquiring the lymph node target image based on the three-dimensional region growing algorithm, the seed points are manually selected according to the report given by the doctor.

[0110] In this embodiment, Figure 4-1 and Figure 4-2 As shown, the corresponding lymph node target image is extracted using the aforementioned second maximum likelihood function, where: Figure 4-1 is the partial lymph node image corresponding to the initial axillary blood vessel-lymph node image, Figure 4-2 The target image of lymph nodes obtained after segmentation.

[0111] S4. Based on the vascular target image and lymph node target image of the preset time series, a vascular-lymph node TIC curve diagram is drawn by calculating the grayscale mean values ​​corresponding to the images.

[0112] Based on the above step S4, for drawing the blood vessel-lymph node TIC curve diagram, it should be noted that:

[0113] Based on the vascular target image and lymph node target image obtained by contrast agent treatment displayed in the DCE-MRI image display window, the vascular target image and lymph node target image corresponding to multiple preset time points are obtained; for the obtained vascular target image and lymph node target image, the grayscale mean of the part of interest in the image is calculated to draw a vascular-lymph node TIC curve graph.

[0114] In this embodiment, the aforementioned multiple preset time points may be time points at the same time interval.

[0115] In this embodiment, the blood vessel-lymph node TIC curve is drawn as follows: Figure 5-1 and Figure 5-2 As shown, Figure 5-1 This is the blood vessel-lymph node TIC curve corresponding to lymph node metastasis. Figure 5-2 This is the blood vessel-lymph node TIC curve corresponding to the absence of lymph node metastasis.

[0116] In this embodiment, based on the above Figure 5-1 and Figure 5-2 , it should be noted that:

[0117] If lymph nodes metastasize, the corresponding vascular TIC curve is more similar to the lymph node TIC curve;

[0118] If lymph nodes have not metastasized, the corresponding vascular TIC curve will be less similar to the lymph node TIC curve.

[0119] In this embodiment, based on the above Figure 5-1 and Figure 5-2 It should also be noted that:

[0120] The foregoing Figure 5-1 The situation shown is only one of many situations in which lymph node metastasis occurs, and the above Figure 5-2 The situation shown is only one of many situations in which lymph node metastases have not occurred.

[0121] S5. Determine the metastasis of axillary lymph nodes of breast cancer by obtaining the mean, standard deviation and median of the TIC curve difference.

[0122] In this embodiment, the TIC curve difference is: the difference between the area under the vascular TIC curve and the area under the lymph node TIC curve in the vascular-lymph node TIC curve diagram.

[0123] Based on the above step S5, for the judgment and prediction of axillary lymph node metastasis of breast cancer, it should be noted that:

[0124] The smaller the mean, standard deviation and median of the TIC curve difference obtained based on the blood vessel-lymph node TIC curve diagram, the greater the probability of axillary lymph node metastasis of breast cancer. Specifically, the lymph node metastasis prediction table is shown in Table 1.

[0125] Table 1. Prediction of lymph node metastasis

[0126] TIC curve difference mean TIC curve difference standard deviation TIC curve difference median Transferred group 1294.17628638456 1048.27557323092 1009.78888888889 No transfer group 2585.75875486138 1616.92867404808 2464.20535020535

[0127] Based on the above Table 1, it should be noted that what is shown in Table 1 is only one of the multiple situations of lymph node metastasis and non-metastasis.

[0128] Based on the above Table 1, it should be noted that, based on Table 1, it can be seen that the mean, standard deviation and median of the TIC curve difference corresponding to when the lymph nodes have metastasized are significantly smaller than the mean, standard deviation and median of the TIC curve difference corresponding to when the lymph nodes have not metastasized.

[0129] The image processing method in Example 2 discloses that a Hessian matrix is ​​constructed based on an initial axillary blood vessel-lymph node image, a blood vessel target image is extracted based on the Hessian matrix, and a lymph node target image is extracted based on the Hessian matrix through a three-dimensional region growing algorithm, so as to achieve accurate three-dimensional segmentation of axillary blood vessels and lymph nodes, and a blood vessel-lymph node TIC curve is drawn based on the acquired blood vessel target image and lymph node image to judge and predict the metastasis of axillary lymph nodes of breast cancer.

[0130] Embodiment 3 provides an image processing system based on axillary blood vessel-lymph node DCE-MRI images, such as Figure 6 As shown, specifically including:

[0131] The image preprocessing module is specifically used to preprocess the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtain the Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image.

[0132] The blood vessel image acquisition module is specifically used to construct a first maximum likelihood function based on the Hessian matrix, and extract a blood vessel target image of a preset time series based on the first maximum likelihood function.

[0133] The lymph node image acquisition module is specifically used to construct a second maximum likelihood function based on the Hessian matrix, determine the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extract the lymph node target image of a preset time series through a three-dimensional region growing algorithm.

[0134] The TIC image acquisition module is specifically used to draw a blood vessel-lymph node TIC curve diagram by calculating the grayscale mean value corresponding to the image based on the blood vessel target image and the lymph node target image of the preset time series.

[0135] Based on the image processing system described in Example 3, through the image preprocessing module, the blood vessel image acquisition module, the lymph node image acquisition module and the TIC image acquisition module, accurate three-dimensional segmentation of axillary blood vessels and lymph nodes can be achieved, and a blood vessel-lymph node TIC curve diagram can be drawn based on the acquired blood vessel target image and lymph node image to judge and predict the metastasis of axillary lymph nodes of breast cancer.

[0136] Embodiment 4 provides an electronic device, such as Figure 7 As shown, specifically including:

[0137] one or more processors;

[0138] A storage device is used to store one or more programs. When the one or more programs are executed by one or more processors, the one or more processors implement the method described in any one of the first and second embodiments.

[0139] It should be noted that in the claims, any figure marks between brackets should not be understood as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "one" or "an" preceding a component does not exclude the presence of multiple such components. In addition, it should be noted that in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" and the like refers to the specific features, structures, materials or characteristics described in conjunction with the embodiment or example included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0140] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments after knowing the basic creative concept. Therefore, the claims should be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0141] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include these modifications and variations.

Claims

1. An image processing method based on axillary blood vessel-lymph node DCE-MRI image, characterized in that: The image processing method comprises: S1, preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image; The preprocessing includes: adjusting the window level for the initial axillary blood vessel-lymph node image; The initial axillary blood vessel-lymph node image is an image processed with a contrast agent, including a blood vessel target image and a lymph node target image; S2. constructing a first maximum likelihood function based on the Hessian matrix, and extracting a blood vessel target image of a preset time series based on the first maximum likelihood function; The first maximum likelihood function is: in, , , is a parameter that controls the sensitivity of the first maximum likelihood function calculation process; in, ； , , is the eigenvalue of the Hessian matrix; S3, constructing a second maximum likelihood function based on the Hessian matrix, determining the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extracting the lymph node target image of a preset time series by a three-dimensional region growing algorithm; The second maximum likelihood function is: in, , , Parameters for controlling the sensitivity of the second maximum likelihood function calculation process; S4, based on the vascular target image and the lymph node target image of the preset time series, drawing a vascular-lymph node TIC curve diagram by calculating the grayscale mean value corresponding to the image; The blood vessel-lymph node TIC curve diagram includes a blood vessel TIC curve and a lymph node TIC curve; S5. Determine the axillary lymph node metastasis of breast cancer by obtaining the mean, standard deviation and median of the TIC curve area difference; The TIC curve area difference is: the difference between the area under the vascular TIC curve and the area under the lymph node TIC curve in the vascular-lymph node TIC curve diagram.

2. The image processing method according to claim 1, characterized in that: The Hessian matrix is ​​obtained based on the preprocessed initial axillary blood vessel-lymph node image in S1, including: Get the eigenvalues ​​of the Hessian matrix , , , and sort the acquired eigenvalues ​​based on the preset eigenvalue sorting condition; The preset characteristic value sorting condition is: ; The Hessian matrix is: in, , , express , , Three directions, It means to find the derivative.

3. The image processing method according to claim 1, characterized in that: The step of determining the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function in S3 includes: Based on the initial axillary blood vessel-lymph node image, the layer where the lymph nodes are located is preliminarily extracted through the second maximum likelihood function; Taking the initially extracted layer as the central layer, one layer is obtained above and below the central layer, and based on the central layer and the layers above and below the central layer, the layer with the brightest grayscale value is obtained; and the location of the lymph node is determined based on the layer with the brightest grayscale value.

4. The image processing method according to claim 1, characterized in that: The step of extracting lymph node target images of a preset time series by using a three-dimensional region growing algorithm in S3 includes: S3-1, calculating the corresponding pixel difference values ​​for the initial pixel and the surrounding adjacent pixels, and establishing a model in which all pixel values ​​are 0; S3-2, according to the preset pixel difference threshold and the absolute value of the pixel difference in step S3-1, the corresponding pixel value in the established model is set to 1; S3-3. For the pixel points in the model whose pixel values ​​are set to 1, repeat the above steps S3-1 and S3-2 until the absolute value of the pixel value difference is greater than the threshold.

5. The image processing method according to claim 1, characterized in that: The step S4, based on the vessel target image and the lymph node target image in the preset time series, draws a vessel-lymph node TIC curve diagram by calculating the grayscale mean corresponding to the image, including: Based on the vascular target image and the lymph node target image obtained after contrast agent treatment, the grayscale mean values ​​corresponding to multiple preset time points are obtained for the area to be studied in the image; The multiple preset time points are time points with the same time interval.

6. An image processing system based on axillary blood vessel-lymph node DCE-MRI images, characterized in that: The image processing system comprises: An image preprocessing module, specifically used for preprocessing the initial axillary blood vessel-lymph node image in the DCE-MRI image display window, and obtaining a Hessian matrix based on the preprocessed initial axillary blood vessel-lymph node image; A blood vessel image acquisition module, specifically used to construct a first maximum likelihood function based on the Hessian matrix, and extract a blood vessel target image of a preset time series based on the first maximum likelihood function; The first maximum likelihood function is: in, , , is a parameter that controls the sensitivity of the first maximum likelihood function calculation process; in, ； , , is the eigenvalue of the Hessian matrix; A lymph node image acquisition module, specifically used to construct a second maximum likelihood function based on the Hessian matrix, determine the location of the lymph nodes in the initial axillary blood vessel-lymph node image based on the second maximum likelihood function, and extract lymph node target images of a preset time series through a three-dimensional region growing algorithm; The second maximum likelihood function is: in, , , Parameters for controlling the sensitivity of the second maximum likelihood function calculation process; A TIC image acquisition module is specifically used to draw a blood vessel-lymph node TIC curve diagram by calculating the grayscale mean value corresponding to the image based on the blood vessel target image and the lymph node target image of the preset time series; The image processing system determines the axillary lymph node metastasis of breast cancer by obtaining the mean, standard deviation and median of the TIC curve area difference; The TIC curve area difference is: the difference between the area under the vascular TIC curve and the area under the lymph node TIC curve in the vascular-lymph node TIC curve diagram.

7. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the method as described in any one of claims 1 to 5.

Images