Neurovascular tumor identification method and system based on multi-modal image fusion

Abstract

The invention relates to the technical field of medical image processing, and discloses a neurovascular tumor recognition method and system based on multi-modal image fusion, and the system comprises a multi-dimensional collection module and an intelligent recognition module. According to the neural vascular tumor identification method and system based on multi-modal image fusion, PET / CT images and MRI images of a tumor part are obtained through a multi-dimensional acquisition module and are classified to form an image set, and an intelligent identification module carries out gray scale range standardization processing on the images of different modalities according to the same time point; according to the method, the difference between images collected by different devices or time points is eliminated, the images are registered and aligned to a unified three-dimensional coordinate system, the visual fusion registration capacity is high, and an intelligent recognition module analyzes the horizontal diameter and the vertical diameter of tumor tissue and the growth state of the tumor tissue according to the three-dimensional coordinate system, generates a tumor data set and a compression index and displays the tumor data set and the compression index. The compression degree of the tumor tissue on the blood vessel wall is evaluated, corresponding treatment suggestions are output, and the intelligent recognition and diagnosis effect is good.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a multi-modal image fusion neurovascular tumor recognition method and system. Background Art

[0002] Neurovascular tumors are a type of tumor that originates from the cerebrovascular system, including cerebral aneurysms, cerebral arteriovenous malformations, and cavernous hemangiomas. These tumors usually occur on the walls of blood vessels in the brain or spinal cord, which may compress surrounding neural structures and cause a series of symptoms, such as headaches, epileptic seizures, and limb weakness. Due to its sensitive location and potential risk of bleeding, timely and accurate diagnosis is crucial for the treatment and prognosis of patients. The diagnosis and treatment of neurovascular tumors require highly accurate medical imaging technology support. CT is one of the preferred methods for screening intracranial lesions. It can quickly obtain high-resolution cross-sectional images, clearly display the anatomical structure of the skull and cerebral blood vessels, and is highly sensitive to acute cerebrovascular events such as cerebral hemorrhage and cerebral infarction. MRI uses magnetic fields and radiofrequency pulses to generate detailed soft tissue images without radiation risk. Multi-parameter imaging helps to comprehensively evaluate tumor characteristics and is particularly suitable for the diagnosis of neurological diseases. DSA can accurately diagnose symptoms such as cerebrovascular stenosis, aneurysms, and vascular malformations by injecting contrast agents and observing vascular morphology and blood flow in real time. In the identification of neurovascular tumors, each medical imaging modality has unique advantages and limitations, and the appropriate imaging method must be carefully selected to achieve the best diagnostic effect.

[0003] At present, the image registration of traditional multimodal image fusion neurovascular tumor recognition methods is difficult. The imaging principles and presentation forms of medical images of different modalities are also different. There are also differences in spatial resolution, grayscale value, etc. There is a lack of standardized processes, and it is difficult to align the granularity of different parameters. The accuracy of tumor recognition is low, which increases the difficulty of tumor diagnosis. Summary of the invention

[0004] 1. Technical issues to be resolved In view of the shortcomings of the existing technology, the present invention provides a multimodal image fusion neurovascular tumor recognition method and system, which has the advantages of strong visual fusion registration capability and good intelligent recognition and diagnosis effect, and solves the problems of difficult image registration and low tumor recognition accuracy in traditional multimodal image fusion neurovascular tumor recognition methods.

[0005] (II) Technical solution To achieve the above object, the present invention provides the following technical solution: a multimodal image fusion neurovascular tumor identification method, comprising the following steps: Step 1: Connect the PET / CT scanner and the magnetic resonance imaging device through the network to obtain the PET / CT images and MRI images of the tumor site, and classify them into a PET / CT image set and a MRI image set; Step 2: Based on the PET / CT image set and the MRI image set, the grayscale range of the images of different modalities is standardized at the same time point, and they are aligned to a unified three-dimensional coordinate system; Step 3: Analyze the horizontal and vertical diameters of the tumor tissue according to the three-dimensional coordinate system and generate the corresponding tumor data set ; Step 4: Based on the PET / CT image set, MRI image set and tumor data set , analyze the growth status of tumor tissue and generate the corresponding compression index ; Step 5: Set a fixed range of compression thresholds , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall and output corresponding treatment recommendations.

[0006] Preferably, in step 1, the expression of the PET / CT image set is , to Indicates the first to the A group of PET / CT images, each group of PET / CT images includes multiple cross-sectional images, and the PET / CT images contain tumor location, morphology, and metabolic tumor volume information. Indicates the time point at which the PET / CT image was acquired.

[0007] Preferably, in step 1, the expression of the MRI image set is , to Indicates the first to the A group of MRI images, each group of MRI images includes T1-weighted images and T2-weighted images. MRI images contain the signal intensity of bones, spinal cord, muscles and tumor tissues. Indicates the time point at which the MRI image was acquired.

[0008] Preferably, in step 2, the standardization process is as follows: According to the PET / CT image set and the MRI image set, the first PET / CT images and A group of MRI images, among which A set of PET / CT images includes multiple slice images. The MRI images included T1-weighted images and T2-weighted images; Use image processing software to obtain PET / CT images and The grayscale values ​​of all pixels in the group MRI images; Extract For any section image in the group PET / CT images, mark the pixel grayscale value of the section image as , to Indicates that in the slice image, the first to the The gray value of each pixel; Calculate the grayscale mean of the cross-section image Grayscale standard deviation , and its calculation formula is as follows: ; ; In the formula, Represents the total number of pixels in the cross-section image. Indicates that in the cross-section image, The gray value of a pixel, , Indicates that the grayscale standard deviation of the cross-section image is calculated according to the standard deviation formula ; Standardization is performed according to the Z-Score normalization formula, which is expressed as follows: ; In the formula, Indicates that in the cross-section image, The gray value of the pixel after standardization; Standardize the PET / CT images and The grayscale values ​​of all pixels in the group MRI image.

[0009] Preferably, in step 2, the registration and alignment process is as follows: According to the standardized A set of MRI images was used to establish a three-dimensional coordinate system, and then the standardized Substitute the PET / CT images into the three-dimensional coordinate system, where The first T1-weighted image is used to highlight the longitudinal relaxation difference of tumor tissue. The first T2-weighted image is used to highlight the transverse relaxation difference of tumor tissue. A group of PET / CT images is used to provide vertical section information of tumor tissue.

[0010] Preferably, in step 3, the tumor data set The calculation process is as follows: Use image processing software to mark the two endpoints of the tumor tissue farthest apart in the horizontal direction in the three-dimensional coordinate system. and , and the two endpoints of the tumor tissue that are farthest apart in the vertical direction and , where the endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of ; ; In the formula, Indicates that the horizontal diameter of the tumor tissue is calculated according to the Euclidean distance formula. It means that the vertical diameter of the tumor tissue is calculated according to the Euclidean distance formula.

[0011] Preferably, in step 4, the compression index The calculation process is as follows: According to the tumor data set , the horizontal diameter of the tumor tissue is marked as , the vertical diameter of the tumor tissue is marked as ; Based on the PET / CT image collection, the metabolic tumor volume is labeled as ; Based on the MRI image collection, the signal intensity of the tumor tissue is marked as ; ; In the formula, Represents the standard value for measuring the horizontal diameter of tumor tissue. represents the weight of the ratio of the horizontal diameter to the standard value, It represents the standard value for measuring the vertical diameter of tumor tissue. The weight representing the ratio of the vertical diameter to the standard value, represents the standard value for measuring metabolic tumor volume, represents the weight of the ratio of metabolic tumor volume to the standard value, It represents the standard value for measuring the signal intensity of tumor tissue. represents the weight of the ratio of tumor tissue signal intensity to the standard value, , , and are constants, and , Indicates according to , , and The compression index of the tumor tissue is calculated.

[0012] Preferably, in step 5, the compression index Equal to the compression threshold It means that tumor cells have already compressed the blood vessel wall and increased vascular permeability. It is recommended to seek medical treatment in time.

[0013] Preferably, in step 5, the compression index Exceeding the oppression threshold When the blood vessels are damaged, it means that the tumor cells have damaged the blood vessel walls and the risk of bleeding is high. It is recommended to intervene with chemotherapy in a timely manner.

[0014] A multi-modal image fusion neurovascular tumor recognition system, including a multi-dimensional acquisition module and an intelligent recognition module; The multi-dimensional acquisition module is connected to a PET / CT scanner and a magnetic resonance imaging device through a network, and is used to acquire a PET / CT image set and an MRI image set, wherein the PET / CT image set includes a PET / CT image of a tumor site, and the MRI image set includes an MRI image set of a tumor site; The intelligent recognition module is composed of a fusion registration unit, a three-dimensional analysis unit and a recognition evaluation unit. The fusion registration unit performs standard processing on the grayscale range of images of different modalities at the same time point according to the PET / CT image set and the MRI image set, and aligns them to a unified three-dimensional coordinate system. The three-dimensional analysis unit analyzes the horizontal diameter and vertical diameter of the tumor tissue according to the three-dimensional coordinate system and generates a corresponding tumor data set. The identification and evaluation unit is based on the PET / CT image set, the MRI image set and the tumor data set. , analyze the growth status of tumor tissue and generate the corresponding compression index The identification and evaluation unit is provided with a fixed range of compression thresholds. , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall and output corresponding treatment recommendations.

[0015] Compared with the prior art, the present invention provides a multi-modal image fusion neurovascular tumor identification method and system, which has the following beneficial effects: The present invention connects the PET / CT scanner and the magnetic resonance imaging device through a multi-dimensional acquisition module network, obtains PET / CT images and MRI images of the tumor site, and classifies them into a PET / CT image set and an MRI image set, so as to comprehensively and accurately obtain detailed information of the tumor. The intelligent recognition module performs standardized processing on the grayscale range of images of different modalities at the same time point according to the PET / CT image set and the MRI image set, eliminates the differences between images acquired by different devices or time points, and aligns them to a unified three-dimensional coordinate system, so that the grayscale values ​​of different tissues and organs are comparable, and various information can correspond accurately, thereby providing more accurate tumor positioning and measurement, and having strong visualization fusion registration capability.

[0016] The present invention uses an intelligent recognition module to analyze the horizontal diameter and vertical diameter of tumor tissue according to a three-dimensional coordinate system and generates a corresponding tumor data set. , realizing the automation and intelligence of tumor identification, improving work efficiency and accuracy, and then based on the PET / CT image set, MRI image set and tumor data set , analyze the growth status of tumor tissue and generate the corresponding compression index , quantify the degree of compression of the tumor on the blood vessel wall, the intelligent recognition module sets a fixed range of compression thresholds , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall, compression index Included in compression threshold When the tumor cells have compressed the blood vessel wall, the vascular permeability has increased, and timely medical treatment is recommended. Exceeding the oppression threshold When it is detected, it means that tumor cells have damaged the blood vessel wall and the risk of bleeding is high. It is recommended to intervene with chemotherapy in time. Intelligent identification and diagnosis are more effective. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a step diagram of the method of the present invention; Figure 2 Schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The image registration of the traditional multimodal image fusion neurovascular tumor recognition method is difficult, the imaging principles and forms of medical images of different modalities are different, there are also differences in spatial resolution, grayscale value, etc., lack of standardized processes, and it is difficult to align the granularity of different parameters. The tumor recognition accuracy is low, which increases the difficulty of tumor diagnosis. Therefore, a multimodal image fusion neurovascular tumor recognition method and system are provided. Please refer to Figure 1 , a multimodal image fusion neurovascular tumor recognition method, comprising the following steps: Step 1: Connect the PET / CT scanner and the magnetic resonance imaging device through the network to obtain the PET / CT images and MRI images of the tumor site, and classify them into a PET / CT image set and an MRI image set to comprehensively and accurately obtain detailed information about the tumor, which will help to improve the accuracy of tumor identification and diagnosis in the future; The expression of PET / CT image set is , to Indicates the first to the A group of PET / CT images, each group of PET / CT images includes multiple cross-sectional images, and the PET / CT images contain tumor location, morphology, and metabolic tumor volume information. Indicates the time point at which the PET / CT image was acquired; The expression of MRI image set is , to Indicates the first to the A group of MRI images, each group of MRI images includes T1-weighted images and T2-weighted images. MRI images contain the signal intensity of bones, spinal cord, muscles and tumor tissues. represents the time point at which the MRI image was acquired; Step 2: Based on the PET / CT image set and the MRI image set, the grayscale range of the images of different modalities is standardized at the same time point to eliminate the differences between the images collected by different devices or time points, and they are aligned to a unified three-dimensional coordinate system to make subsequent analysis more reliable and consistent; The standardized processing flow is as follows: According to the PET / CT image set and the MRI image set, the first PET / CT images and A group of MRI images, among which A set of PET / CT images includes multiple slice images. The MRI images included T1-weighted images and T2-weighted images; Use image processing software to obtain PET / CT images and The grayscale values ​​of all pixels in the group MRI images; Extract For any section image in the group PET / CT images, mark the pixel grayscale value of the section image as , to Indicates that in the slice image, the first to the The gray value of each pixel; Calculate the grayscale mean of the cross-section image Grayscale standard deviation , and its calculation formula is as follows: ; ; In the formula, Represents the total number of pixels in the cross-section image, Indicates that in the cross-section image, The gray value of a pixel, , Indicates that the grayscale standard deviation of the cross-section image is calculated according to the standard deviation formula ; Standardization is performed according to the Z-Score normalization formula, which is expressed as follows: ; In the formula, Indicates that in the cross-section image, The gray value of the pixel after standardization; In the actual processing process, in order to ensure the effect of standardization, preprocessing can be considered before standardization, such as removing some extremely high or low grayscale outliers, so that the grayscale values ​​of different tissues and organs are comparable, which is convenient for subsequent medical image analysis, diagnosis and research; According to the above process, standardize the PET / CT images and The grayscale values ​​of all pixels in the group MRI images; The registration and alignment process is as follows: According to the standardized A set of MRI images was used to establish a three-dimensional coordinate system, and then the standardized Substitute the PET / CT images into the three-dimensional coordinate system, where The first T1-weighted image is used to highlight the longitudinal relaxation difference of tumor tissue. The first T2-weighted image is used to highlight the transverse relaxation difference of tumor tissue. The PET / CT images are used to provide the vertical section information of the tumor tissue. Various information can be accurately matched, thus providing more accurate tumor positioning and measurement, and the visualization fusion registration capability is strong; Step 3: Analyze the horizontal and vertical diameters of the tumor tissue according to the three-dimensional coordinate system and generate the corresponding tumor data set , the calculation process is as follows: Use image processing software to mark the two endpoints of the tumor tissue farthest apart in the horizontal direction in the three-dimensional coordinate system. and , and the two endpoints of the tumor tissue that are farthest apart in the vertical direction and , where the endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of ; ; In the formula, Indicates that the horizontal diameter of the tumor tissue is calculated according to the Euclidean distance formula. It means that the vertical diameter of the tumor tissue can be calculated according to the Euclidean distance formula, which realizes the automation and intelligence of tumor identification and improves work efficiency and accuracy; Step 4: Based on the PET / CT image set, MRI image set and tumor data set , analyze the growth status of tumor tissue and generate the corresponding compression index The calculation process is as follows: According to the tumor data set , the horizontal diameter of the tumor tissue is marked as , the vertical diameter of the tumor tissue is marked as ; Based on the PET / CT image collection, the metabolic tumor volume is labeled as ; Based on the MRI image collection, the signal intensity of the tumor tissue is marked as ; ; In the formula, Represents the standard value for measuring the horizontal diameter of tumor tissue. represents the weight of the ratio of the horizontal diameter to the standard value, It represents the standard value for measuring the vertical diameter of tumor tissue. The weight representing the ratio of the vertical diameter to the standard value, represents the standard value for measuring metabolic tumor volume, represents the weight of the ratio of metabolic tumor volume to the standard value, It represents the standard value for measuring the signal intensity of tumor tissue. represents the weight of the ratio of tumor tissue signal intensity to the standard value, , , and are constants, and , Indicates according to , , and Weights are used to calculate the compression index of tumor tissue, quantify the degree of compression of the tumor on the blood vessel wall, and provide a scientific basis for clinical decision-making; Step 5: Set a fixed range of compression thresholds , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall, compression index Equal to the compression threshold When the tumor cells have compressed the blood vessel wall, the vascular permeability has increased, and timely medical treatment is recommended. Exceeding the oppression threshold When it is detected, it means that tumor cells have damaged the blood vessel wall and the risk of bleeding is high. It is recommended to intervene with chemotherapy in time. Intelligent identification and diagnosis are more effective.

[0020] See also Figure 2 , a multimodal image fusion neurovascular tumor recognition system, including a multidimensional acquisition module and an intelligent recognition module; The multi-dimensional acquisition module is connected to the PET / CT scanner and the magnetic resonance imaging device through a network, and is used to acquire a PET / CT image set and an MRI image set, wherein the PET / CT image set includes a PET / CT image of a tumor site, and the MRI image set includes an MRI image set of a tumor site; The intelligent recognition module consists of a fusion registration unit, a three-dimensional analysis unit, and a recognition evaluation unit. The fusion registration unit standardizes the grayscale range of images of different modalities at the same time point based on the PET / CT image set and the MRI image set, and aligns them to a unified three-dimensional coordinate system. It has strong visualization fusion registration capabilities. The three-dimensional analysis unit analyzes the horizontal and vertical diameters of tumor tissues based on the three-dimensional coordinate system and generates the corresponding tumor data set. The recognition and evaluation unit is based on the PET / CT image set, the MRI image set and the tumor data set. , analyze the growth status of tumor tissue and generate the corresponding compression index , the recognition evaluation unit is set with a fixed range of compression thresholds , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall, and output corresponding treatment recommendations, assisting medical staff to make fast and accurate treatment decisions, with intelligent identification and diagnosis effects.

[0021] Example 1: In this example, a patient with a neurovascular tumor in the brain was selected as an experimental subject. After testing, in the three-dimensional coordinate system of the patient's tumor site, the two endpoints of the tumor tissue farthest apart in the horizontal direction were and , the two endpoints of the tumor tissue that are farthest apart in the vertical direction are and , endpoint The coordinates are (3, 4, 5), and the endpoints The coordinates are (7, 8, 6), and the endpoints The coordinates are (2, 1, 3), and the endpoints The coordinates of the patient's tumor data set are (6, 9, 10). The calculation process is as follows: ; ; In the formula, according to the Euclidean distance formula, the horizontal diameter of the tumor tissue is calculated to be 5.74, and according to the Euclidean distance formula, the vertical diameter of the tumor tissue is calculated to be 11.36.

[0022] Example 2: In this example, a patient with a neurovascular tumor growing in the neck was selected as the experimental subject. After testing, the horizontal diameter of the patient's tumor tissue was 5 cm, the vertical diameter was 4 cm, the metabolic tumor volume was 30 cm³, the signal intensity was 80 units, and the compression index of the patient was The calculation process is as follows: ; ; ; In the formula, the standard value for measuring the horizontal diameter of tumor tissue is 6 cm, and the weight of the ratio of the horizontal diameter to the standard value is 0.25. The standard value for measuring the vertical diameter of tumor tissue is 5 cm, and the weight of the ratio of the vertical diameter to the standard value is 0.25. The standard value for measuring the metabolic tumor volume is 35 cm³, and the weight of the ratio of the metabolic tumor volume to the standard value is 0.25. The standard value for measuring the signal intensity of tumor tissue is 90 units, and the weight of the ratio of the signal intensity of tumor tissue to the standard value is 0.25. , , and are constants, and ,according to , , and The compression index of the tumor tissue was calculated to be approximately 0.85.

[0023] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multimodal image fusion neurovascular tumor recognition method, characterized by: The following steps are involved: Step 1: Connect the PET / CT scanner and the magnetic resonance imaging device through the network to obtain the PET / CT images and MRI images of the tumor site, and classify them into a PET / CT image set and a MRI image set; Step 2: Based on the PET / CT image set and the MRI image set, the grayscale range of the images of different modalities is standardized at the same time point, and they are aligned to a unified three-dimensional coordinate system; Step 3: Analyze the horizontal and vertical diameters of the tumor tissue according to the three-dimensional coordinate system and generate the corresponding tumor data set ; Step 4: Based on the PET / CT image set, MRI image set and tumor data set , analyze the growth status of tumor tissue and generate the corresponding compression index ; Step 5: Set a fixed range of compression thresholds , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall and output corresponding treatment recommendations.

2. The multimodal image fusion neurovascular tumor recognition method according to claim 1, characterized in that: In step 1, the expression of the PET / CT image set is: , to Indicates the first to the A group of PET / CT images, each group of PET / CT images includes multiple cross-sectional images, and the PET / CT images contain tumor location, morphology, and metabolic tumor volume information. Indicates the time point at which the PET / CT image was acquired.

3. The multimodal image fusion neurovascular tumor recognition method according to claim 2, characterized in that: In step 1, the expression of the MRI image set is: , to Indicates the first to the A group of MRI images, each group of MRI images includes T1-weighted images and T2-weighted images. MRI images contain the signal intensity of bones, spinal cord, muscles and tumor tissues. Indicates the time point at which the MRI image was acquired.

4. The multimodal image fusion neurovascular tumor recognition method according to claim 3, characterized in that: In step 2, the standardization process is as follows: According to the PET / CT image set and the MRI image set, the first PET / CT images and A group of MRI images, among which A set of PET / CT images includes multiple slice images. The MRI images included T1-weighted images and T2-weighted images; Use image processing software to obtain PET / CT images and The grayscale values ​​of all pixels in the group MRI images; Extract For any section image in the group PET / CT images, mark the pixel grayscale value of the section image as , to Indicates that in the slice image, the first to the The gray value of each pixel; Calculate the grayscale mean of the cross-section image Grayscale standard deviation , and its calculation formula is as follows: ； ； In the formula, Represents the total number of pixels in the cross-section image. Indicates that in the cross-section image, The gray value of a pixel, , Indicates that the grayscale standard deviation of the cross-section image is calculated according to the standard deviation formula ; Standardization is performed according to the Z-Score normalization formula, which is expressed as follows: ； In the formula, Indicates that in the cross-section image, The gray value of the pixel after standardization; Standardize the PET / CT images and The grayscale values ​​of all pixels in the group MRI image.

5. The multimodal image fusion neurovascular tumor recognition method according to claim 4, characterized in that: In step 2, the registration and alignment process is as follows: According to the standardized A set of MRI images was used to establish a three-dimensional coordinate system, and then the standardized Substitute the PET / CT images into the three-dimensional coordinate system, where The first T1-weighted image is used to highlight the longitudinal relaxation difference of tumor tissue. The first T2-weighted image is used to highlight the transverse relaxation difference of tumor tissue. A group of PET / CT images is used to provide vertical section information of tumor tissue.

6. The multimodal image fusion neurovascular tumor recognition method according to claim 5, characterized in that: In step 3, the tumor data set The calculation process is as follows: Use image processing software to mark the two endpoints of the tumor tissue farthest apart in the horizontal direction in the three-dimensional coordinate system. and , and the two endpoints of the tumor tissue that are farthest apart in the vertical direction and , where the endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of , endpoint The coordinates of ; ； In the formula, Indicates that the horizontal diameter of the tumor tissue is calculated according to the Euclidean distance formula. It means that the vertical diameter of the tumor tissue is calculated according to the Euclidean distance formula.

7. The multimodal image fusion neurovascular tumor recognition method according to claim 6, characterized in that: In step 4, the compression index The calculation process is as follows: According to the tumor data set , the horizontal diameter of the tumor tissue is marked as , the vertical diameter of the tumor tissue is marked as ; Based on the PET / CT image collection, the metabolic tumor volume is labeled as ; Based on the MRI image set, the signal intensity of the tumor tissue is marked as ; ； In the formula, Represents the standard value for measuring the horizontal diameter of tumor tissue. represents the weight of the ratio of the horizontal diameter to the standard value, It represents the standard value for measuring the vertical diameter of tumor tissue. The weight representing the ratio of the vertical diameter to the standard value, represents the standard value for measuring metabolic tumor volume, represents the weight of the ratio of metabolic tumor volume to the standard value, It represents the standard value for measuring the signal intensity of tumor tissue. represents the weight of the ratio of tumor tissue signal intensity to the standard value, , , and are constants, and , Indicates according to , , and The compression index of the tumor tissue is calculated.

8. The multimodal image fusion neurovascular tumor recognition method according to claim 7, characterized in that: Step 5: Compression Index Equal to the compression threshold It means that tumor cells have already compressed the blood vessel wall and increased vascular permeability. It is recommended to seek medical treatment in time.

9. The multimodal image fusion neurovascular tumor recognition method according to claim 8, characterized in that: Step 5: Compression Index Exceeding the oppression threshold When the blood vessels are damaged, it means that the tumor cells have damaged the blood vessel walls and the risk of bleeding is high. It is recommended to intervene with chemotherapy in a timely manner.

10. A multi-modal image fusion neurovascular tumor recognition system, using a multi-modal image fusion neurovascular tumor recognition method according to any one of claims 1 to 9, characterized in that: The system includes a multi-dimensional acquisition module and an intelligent recognition module; The multi-dimensional acquisition module is connected to a PET / CT scanner and a magnetic resonance imaging device through a network, and is used to acquire a PET / CT image set and an MRI image set, wherein the PET / CT image set includes a PET / CT image of a tumor site, and the MRI image set includes an MRI image set of a tumor site; The intelligent recognition module is composed of a fusion registration unit, a three-dimensional analysis unit and a recognition evaluation unit. The fusion registration unit performs standard processing on the grayscale range of images of different modalities at the same time point according to the PET / CT image set and the MRI image set, and aligns them to a unified three-dimensional coordinate system. The three-dimensional analysis unit analyzes the horizontal diameter and vertical diameter of the tumor tissue according to the three-dimensional coordinate system and generates a corresponding tumor data set. The identification and evaluation unit is based on the PET / CT image set, the MRI image set and the tumor data set. , analyze the growth status of tumor tissue and generate the corresponding compression index The identification and evaluation unit is provided with a fixed range of compression thresholds. , combined with the compression index , evaluate the degree of compression of tumor tissue on the blood vessel wall and output corresponding treatment recommendations.

Images