Image-based severe pancreatitis assessment method and system

Abstract

The invention provides an image-based severe pancreatitis assessment method and system. The method comprises the following steps: sequentially carrying out image size adjustment, Gaussian filtering processing and image enhancement processing on an abdominal nuclear magnetic resonance image; performing clustering processing and sparse denoising processing on the abdomen CT image in sequence; performing convolution fusion processing on the nuclear magnetic processing image and the CT processing image; performing region-of-interest segmentation on the fused image based on the pancreas template image; and inputting the segmented feature map into the final evaluation model for evaluation so as to output a severe pancreatitis evaluation result, the image definition and the image quality can be effectively improved, artifact noise and the like existing in the CT image can be removed at the same time, and the evaluation accuracy and effectiveness can be further improved at the same time.

Description

Technical Field

[0001] The present invention belongs to the technical field of severe pancreatitis assessment, and specifically relates to an image-based severe pancreatitis assessment method and system. Background Art

[0002] Severe pancreatitis is a special type of pancreatitis. It is more serious, has more complications, and has a high mortality rate. Patients with severe pancreatitis often have systemic and local complications, such as systemic inflammatory reactions, and some patients suffer from septic shock. For patients with severe pancreatitis, first consider fasting and water, gastrointestinal decompression, and intravenous use of enzyme inhibitors and digestive juice secretion inhibitors for further treatment.

[0003] Currently, the assessment of severe pancreatitis in patients is mostly performed by taking abdominal CT images of the patients and combining them with the doctor's experience. However, the existing CT images are generally low-dose CT images, which result in insufficient image clarity and low image quality. There is a lot of noise in the image, which in turn affects the assessment of severe pancreatitis and reduces the accuracy of the assessment. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an image-based severe pancreatitis assessment method and system for solving the technical problems in the prior art.

[0005] On the one hand, the present invention provides the following technical solution, a method for evaluating severe pancreatitis based on imaging, comprising: Acquire an abdominal CT image and an abdominal MRI image of the patient, and sequentially perform image size adjustment, Gaussian filtering, and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; Performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; Acquire a pancreatic template image, and perform region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; A training image is obtained, and the training image is input into a preset evaluation model for training to obtain a final evaluation model, and the segmentation feature map is input into the final evaluation model for evaluation to output a severe pancreatitis evaluation result.

[0006] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention first obtains an abdominal CT image and an abdominal nuclear magnetic resonance image of a patient, and sequentially performs image size adjustment, Gaussian filtering and image enhancement processing on the abdominal nuclear magnetic resonance image to obtain a nuclear magnetic processing image; then the abdominal CT image is sequentially subjected to clustering processing and sparse denoising processing to obtain a CT processing image; then the nuclear magnetic processing image and the CT processing image are subjected to convolution fusion processing to obtain a fused image; then a pancreatic template image is obtained, and the fused image is segmented into a region of interest based on the pancreatic template image to obtain a segmentation feature map; finally, a training image is obtained, and the training image is input into a preset evaluation model for training to obtain a final evaluation model, and the segmentation feature map is input into the final evaluation model for evaluation to output a severe pancreatitis evaluation result. The present invention first performs enhancement and denoising processing on the image, which can effectively improve the clarity and image quality of the image while removing artifact noise and the like existing in the CT image, and the present invention effectively fuses the nuclear magnetic image with the CT image, which can further improve the evaluation accuracy. The present invention effectively segments the abdominal image of the patient and makes the location of the patient's lesion more obvious, thereby improving the accuracy and effectiveness of subsequent model evaluation.

[0007] Preferably, the step of sequentially performing clustering processing and sparse denoising processing on the abdominal CT image to obtain a CT processed image comprises: Dividing the abdominal CT image into blocks to obtain a plurality of image blocks; The K clustering algorithm is used to cluster the image blocks to obtain a number of clusters: Construct the initial feature dictionary and set the initial sparse error, and determine the covariance matrix based on the initial feature field and the initial sparse error : ; In the formula, represents the initial sparse error, represents the initial feature dictionary, express The initial diagonal matrix of order, The diagonal elements in are , ,… , Indicates Variation parameters; The change parameters in the initial diagonal matrix are iteratively adjusted, and a CT processed image is determined based on the adjusted change parameters.

[0008] Preferably, the step of sequentially performing clustering processing and sparse denoising processing on the abdominal CT image to obtain a CT processed image comprises: Iteratively adjust the change parameters in the initial diagonal matrix, and iteratively update the sparse error based on the adjusted change parameters to obtain an iterative error : ; In the formula, , Respectively represent sequence The iterative error after iterations is represents the image blocks in the cluster, Indicates The feature dictionary after iterations, represents the number of column vectors in the feature dictionary, Indicates the change parameter adjustment The covariance matrix after the second step; Based on the iterative error Determine The updated sparse matrix after iterations : ; In the formula, Indicates Feature dictionary after iterations; Based on The updated sparse matrix after iterations Determine the decomposition matrix : ; In the formula, represents the subscript matrix in the feature dictionary, Indicates The column vector in the feature dictionary after iterations; For the decomposition matrix Perform singular value decomposition to obtain a first orthogonal matrix and a second orthogonal matrix, and use the column vector in the first orthogonal matrix as the first The column vector of the feature dictionary after the iteration, the product of the first column vector of the second orthogonal matrix and the first eigenvalue is taken as the The column vector in the coefficient matrix after iterations; Repeat the iterative process until the iteration stop number is met and input the corresponding iterative feature dictionary and iterative coefficient matrix; Based on the iterated feature dictionary And the coefficient matrix after iteration De-noising the image blocks in the cluster to obtain a CT processed image : ; In the formula, is Lagrangian sparse, represents the noisy area in the image block, represents the residual factor, represents the sparse identity matrix.

[0009] Preferably, the step of performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image comprises: Constructing a first Gaussian filter and a plurality of second Gaussian filters in different directions, and inputting the first Gaussian filter and the second Gaussian filter into a deconvolution network for training and updating to obtain a first update filter and a second update filter; Inputting the nuclear magnetic resonance processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering to obtain a first low-frequency nuclear magnetic resonance image and a plurality of second high-frequency nuclear magnetic resonance images; Inputting the CT processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering to obtain a first low-frequency CT image and a plurality of second CT high-frequency images; The first low-frequency nuclear magnetic resonance image is fused with the first low-frequency CT image to obtain a first fused image. , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image ; The first fused image input to the first update filter and the second fused image respectively input into the second update filter for convolution and summation to obtain a number of undetermined fused images; The particle swarm optimization algorithm is used to iteratively determine the optimal iteration frequency of the first update filter and the second update filter, and the undetermined fusion image corresponding to the optimal iteration frequency is used as the fusion image.

[0010] Preferably, the first low-frequency MRI image is fused with the first low-frequency CT image to obtain a first fused image. , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image The steps include: Calculate a first low-frequency fusion weight based on the first low-frequency nuclear magnetic resonance image and the first low-frequency CT image Fusion weight with the second low frequency : ; ; In the formula, , Respectively represent the pixel coordinates of the first low-frequency MRI image and the first low-frequency CT image, , Represent the first and second adjustment parameters respectively, , Represent the first and second fixed parameters respectively, Respectively represent the gradient matrices in the first low-frequency MRI image and the first low-frequency CT image, represents the first Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; Based on the first low frequency fusion weight Fusion weight with the second low frequency The first low-frequency nuclear magnetic resonance image , the first low-frequency CT image Fusion is performed to obtain the first fused image : ; Based on the second high-frequency nuclear magnetic image The second high-frequency CT image Determine the first high frequency filtered image With the second high frequency filtered image : ; ; In the formula, represents the characteristic scale, Indicates scale structural elements, , They represent morphological dilation and erosion operations respectively; Based on the first high frequency filtered image With the second high frequency filtered image Confirm the first high frequency filter value respectively With the second high frequency filter value : ; ; In the formula, , Respectively represent the pixel coordinates of the first high-frequency filtered image and the second high-frequency filtered image, , Respectively represent the third and fourth adjustment parameters, , Respectively represent the third and fourth fixed parameters, Respectively represent the gradient matrices in the first high-frequency filtered image and the second high-frequency filtered image, represents the second Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; Based on the first high frequency filter value With the second high frequency filter value Calculate the first high frequency fusion weight Fusion weight with the second low frequency : ; ; Based on the first high frequency fusion weight Fusion weight with the second low frequency The second high-frequency nuclear magnetic resonance image , Second high frequency CT image Fusion is performed to obtain a second fused image : .

[0011] Preferably, the step of performing region of interest segmentation on the fused image based on the pancreas template image to obtain a segmentation feature map comprises: Storing the pancreas template image in a template image set, and dividing the template image set into a first image set and a second image set according to a preset ratio; Randomly selecting an image in the first image set as an initial pancreatic image, and aligning the remaining images in the first image set to the initial pancreatic image to obtain a plurality of registration and alignment matrices, and calculating an average of the plurality of registration and alignment matrices to obtain an average alignment matrix; All images in the second image set are registered and aligned to the initial pancreas image, and the image in the second image set that best matches the average alignment matrix is ​​selected as the standard image; The fused image is segmented based on the standard image to obtain a segmentation feature map.

[0012] Preferably, the step of segmenting the fused image based on the standard image to obtain a segmentation feature map includes: Aligning the fused image to the standard image to obtain a fused alignment matrix, and calculating an average value of the fused alignment matrix and a plurality of the registration alignment matrices to obtain a fused average alignment matrix; Re-iteratively determining the standard image and the registration alignment process based on the fused average alignment matrix until the iteration termination condition is met, and outputting the final template image; Extract the binary image of the final template image to obtain a template image pair, and calculate the likelihood image of the fused image : ; In the formula, Represents the likelihood image The pixel value of a pixel, Indicates the fused image The pixel value of a pixel, represents the likelihood expectation, represents the likelihood variance; Based on the likelihood image Compared with the template image Determine segmentation feature map : ; In the formula, Represents and operations.

[0013] In a second aspect, the present invention provides the following technical solution: an image-based severe pancreatitis assessment system, the system comprising: An enhancement module is used to obtain an abdominal CT image and an abdominal MRI image of a patient, and to sequentially perform image size adjustment, Gaussian filtering and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; A denoising module, used for performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; A fusion module, used for performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; A segmentation module, used for acquiring a pancreatic template image, and performing region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; The evaluation module is used to obtain training images, input the training images into a preset evaluation model for training to obtain a final evaluation model, input the segmentation feature map into the final evaluation model for evaluation to output a severe pancreatitis evaluation result.

[0014] In a third aspect, the present invention provides the following technical solution: a computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned image-based severe pancreatitis assessment method when executing the computer program.

[0015] In a fourth aspect, the present invention provides the following technical solution: a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned image-based severe pancreatitis assessment method. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A flowchart of an imaging-based severe pancreatitis assessment method provided in Example 1 of the present invention; Figure 2 A structural block diagram of an image-based severe pancreatitis assessment system provided in Embodiment 2 of the present invention; Figure 3 A schematic diagram of the hardware structure of a computer provided in another embodiment of the present invention.

[0018] The embodiments of the present invention will be further described below with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the embodiments of the present invention, and should not be construed as limiting the present invention.

[0020] In the description of the embodiments of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0021] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0022] In the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific circumstances.

[0023] Embodiment 1 In the first embodiment of the present invention, Figure 1 As shown in Figure 1, an imaging-based approach to the evaluation of severe pancreatitis includes: S1, obtaining an abdominal CT image and an abdominal MRI image of a patient, and sequentially performing image size adjustment, Gaussian filtering and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; Specifically, in this step, resizing, Gaussian filtering and image enhancement are all commonly used image processing methods in the prior art, so they are not described in detail here.

[0024] S2, performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; Specifically, step S2 includes: S21, dividing the abdominal CT image into blocks to obtain a plurality of image blocks; Specifically, in this step, the abdominal CT image is segmented into a number of image blocks of the same size according to the size of the abdominal CT image.

[0025] S22, clustering the plurality of image blocks using a K clustering algorithm to obtain a plurality of clusters: Specifically, the K clustering algorithm used here is the K-means clustering algorithm in the prior art, so it will not be described in detail here.

[0026] S23, construct an initial feature dictionary and set an initial sparse error, and determine the covariance matrix based on the initial feature field and the initial sparse error : ; In the formula, represents the initial sparse error, represents the initial feature dictionary, express The initial diagonal matrix of order, The diagonal elements in are , ,… , Indicates Variation parameters; Specifically, the feature field here is specifically a sparse field, and the precise feature dictionary and sparse representation are determined through subsequent sparse erasure. At the same time, in this application, the change parameter is specifically a hyperparameter, which controls the degree of change of each column vector in the feature dictionary.

[0027] S24, iteratively adjusting the change parameters in the initial diagonal matrix, and determining a CT processed image based on the adjusted change parameters; Wherein, the step S24 comprises: S241, iteratively adjusting the change parameters in the initial diagonal matrix, and iteratively updating the sparse error based on the adjusted change parameters to obtain an iterative error : ; In the formula, , Respectively represent sequence The iterative error after iterations is represents the image blocks in the cluster, Indicates The feature dictionary after iterations, represents the number of column vectors in the feature dictionary, Indicates the change parameter adjustment The covariance matrix after .

[0028] S242, based on the iterative error Determine The updated sparse matrix after iterations : ; In the formula, Indicates Feature dictionary after iterations.

[0029] S243, based on The updated sparse matrix after iterations Determine the decomposition matrix : ; In the formula, represents the subscript matrix in the feature dictionary, Indicates The column vector in the feature dictionary after iterations; Specifically, the subscript matrix in the feature dictionary can be determined by the subscript sort value of each element in the feature dictionary. Meanwhile, after several iterations, the number of elements in the feature dictionary will not change, but the element values ​​will change with the iterations.

[0030] S244, decomposing the matrix Perform singular value decomposition to obtain a first orthogonal matrix and a second orthogonal matrix, and use the column vector in the first orthogonal matrix as the first The column vector of the feature dictionary after the iteration, the product of the first column vector of the second orthogonal matrix and the first eigenvalue is taken as the The column vector in the coefficient matrix after iterations; Specifically, assuming that the decomposition matrix is ​​an i×j matrix, after singular value decomposition, an i×i orthogonal matrix and a j×j orthogonal matrix will be obtained. The i×i orthogonal matrix is ​​the first orthogonal matrix, and the j×j orthogonal matrix is ​​the second orthogonal matrix.

[0031] S245, repeat the iterative process until the iteration stop times are met and input the corresponding post-iteration feature dictionary and post-iteration coefficient matrix.

[0032] S246, based on iterated feature dictionary And the coefficient matrix after iteration De-noising the image blocks in the cluster to obtain a CT processed image : ; In the formula, is Lagrangian sparse, represents the noisy area in the image block, represents the residual factor, represents the sparse identity matrix.

[0033] S3, performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; Wherein, the step S3 comprises: S31, constructing a first Gaussian filter and a plurality of second Gaussian filters in different directions, inputting the first Gaussian filter and the second Gaussian filter into a deconvolution network for training and updating, so as to obtain a first update filter and a second update filter; Specifically, in this step, the first Gaussian filter is a low-pass filter, and the second Gaussian filter is a Gaussian directional filter with different angles. In the present application, there is one first Gaussian filter and eight second Gaussian filters, and their angular directions are in an arithmetically progressive relationship.

[0034] S32, inputting the nuclear magnetic resonance processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering, so as to obtain a first low-frequency nuclear magnetic resonance image and a plurality of second high-frequency nuclear magnetic resonance images; Specifically, after passing through the above filters, the images will undergo an adaptive decomposition process to obtain corresponding low-frequency images and high-frequency images.

[0035] S33, inputting the CT processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering processing, so as to obtain a first low-frequency CT image and a plurality of second CT high-frequency images.

[0036] S34, fusing the first low-frequency nuclear magnetic resonance image with the first low-frequency CT image to obtain a first fused image , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image ; Wherein, the step S34 comprises: S341, calculating a first low-frequency fusion weight based on the first low-frequency nuclear magnetic resonance image and the first low-frequency CT image Fusion weight with the second low frequency : ; ; In the formula, , Respectively represent the pixel coordinates of the first low-frequency MRI image and the first low-frequency CT image, , Represent the first and second adjustment parameters respectively, , Represent the first and second fixed parameters respectively, Respectively represent the gradient matrices in the first low-frequency MRI image and the first low-frequency CT image, represents the first Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at .

[0037] S342, based on the first low-frequency fusion weight Fusion weight with the second low frequency The first low-frequency nuclear magnetic resonance image , the first low-frequency CT image Fusion is performed to obtain the first fused image : .

[0038] S343, based on the second high frequency nuclear magnetic resonance image The second high-frequency CT image Determine the first high frequency filtered image With the second high frequency filtered image : ; ; In the formula, represents the characteristic scale, Indicates scale structural elements, , They represent morphological dilation and erosion operations respectively.

[0039] S344: Based on the first high-frequency filtered image With the second high frequency filtered image Confirm the first high frequency filter value respectively With the second high frequency filter value : ; ; In the formula, , Respectively represent the pixel coordinates of the first high-frequency filtered image and the second high-frequency filtered image, , Respectively represent the third and fourth adjustment parameters, , Respectively represent the third and fourth fixed parameters, Respectively represent the gradient matrices in the first high-frequency filtered image and the second high-frequency filtered image, represents the second Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at .

[0040] S345, based on the first high frequency filter value With the second high frequency filter value Calculate the first high frequency fusion weight Fusion weight with the second low frequency : ; .

[0041] S346, based on the first high frequency fusion weight Fusion weight with the second low frequency The second high-frequency nuclear magnetic resonance image , Second high frequency CT image Fusion is performed to obtain a second fused image : ; Specifically, in this process, images processed by the same second Gaussian filter are fused together, so eight second fused images are finally obtained.

[0042] S35, the first fused image input to the first update filter and the second fused image The images are respectively input into the second update filter for convolution and summation to obtain several undetermined fused images.

[0043] S36, using a particle swarm optimization algorithm to iteratively determine the optimal frequency of iteration of the first update filter and the second update filter, and using the undetermined fused image corresponding to the optimal frequency of iteration as the fused image; Specifically, the particle swarm optimization algorithm here specifically sets the position and speed of the particles and calculates the fitness, individual extreme value and global extreme value of each particle, and updates the speed, position, individual extreme value and global extreme value of the particles, and recalculates the fitness and the absolute value of the difference between two adjacent fitness values, and determines whether to execute the iterative process based on the difference between the two and the threshold, until the optimal solution is output. The optimal solution here is the optimal cutoff frequency of the filter.

[0044] S4, acquiring a pancreatic template image, and performing region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; Wherein, the step S4 comprises: S41, storing the pancreas template image into a template image set, and dividing the template image set into a first image set and a second image set according to a preset ratio; Specifically, the first image set is a training template set, the second image set is a test template set, and the preset ratio is 80%, with the training template set accounting for 80%.

[0045] S42. Randomly select an image in the first image set as an initial pancreatic image, and align the remaining images in the first image set to the initial pancreatic image to obtain a plurality of registration and alignment matrices, and calculate an average of the plurality of registration and alignment matrices to obtain an average alignment matrix.

[0046] S43, all images in the second image set are registered and aligned to the initial pancreas image, and an image in the second image set that best matches the average alignment matrix is ​​selected as a standard image.

[0047] S44, segmenting the fused image based on the standard image to obtain a segmentation feature map; Wherein, the step S44 comprises: S441, aligning the fused image to the standard image to obtain a fused alignment matrix, and calculating an average value of the fused alignment matrix and a plurality of the registration alignment matrices to obtain a fused average alignment matrix.

[0048] S442, re-iteratively determining the standard image and the registration alignment process based on the fused average alignment matrix until an iteration termination condition is met, and outputting a final template image; Specifically, by repeating the above steps, the best matching final template image can be output, and the specific position of the pancreas is marked on the final template image.

[0049] S443: extract the binary image of the final template image to obtain a template image pair, and calculate the likelihood image of the fused image. : ; In the formula, Represents the likelihood image The pixel value of a pixel, Indicates the fused image The pixel value of a pixel, represents the likelihood expectation, represents the likelihood variance.

[0050] S444: Based on the likelihood image Compared with the template image Determine segmentation feature map : ; In the formula, Represents and operations.

[0051] S5, obtaining a training image, inputting the training image into a preset evaluation model for training to obtain a final evaluation model, inputting the segmentation feature map into the final evaluation model for evaluation to output a severe pancreatitis evaluation result; Specifically, in this embodiment, the training images are images of patients with severe pancreatitis, and in this embodiment, the preset evaluation model is specifically a Softmax classifier.

[0052] The image-based severe pancreatitis assessment method provided in the first embodiment of the present invention first obtains the patient's abdominal CT image and abdominal nuclear magnetic resonance image, and sequentially performs image size adjustment, Gaussian filtering and image enhancement on the abdominal nuclear magnetic resonance image to obtain a nuclear magnetic processing image; then the abdominal CT image is sequentially clustered and sparsely denoised to obtain a CT processing image; then the nuclear magnetic processing image and the CT processing image are convolutionally fused to obtain a fused image; then a pancreatic template image is obtained, and the fused image is segmented into a region of interest based on the pancreatic template image to obtain a segmentation feature map; finally, a training image is obtained, and the training image is input into a preset evaluation model for training to obtain a final evaluation model, and the segmentation feature map is input into the final evaluation model for evaluation to output a severe pancreatitis assessment result. The present invention first performs enhancement and denoising on the image, which can effectively improve the clarity and image quality of the image while removing artifact noise and the like in the CT image, and the present invention effectively fuses the nuclear magnetic image with the CT image, which can further improve the assessment accuracy. The present invention effectively segments the patient's abdominal image and makes the patient's lesion location more obvious, thereby improving the accuracy and effectiveness of subsequent model evaluation.

[0053] Embodiment 2 like Figure 2As shown, in the second embodiment of the present invention, an image-based severe pancreatitis assessment system is provided, and the system includes: Enhancement module 1 is used to obtain the patient's abdominal CT image and abdominal MRI image, and sequentially perform image size adjustment, Gaussian filtering and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; De-noising module 2, used for performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; A fusion module 3 is used to perform convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; A segmentation module 4 is used to obtain a pancreatic template image, and perform region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; Evaluation module 5, used for acquiring training images, inputting the training images into a preset evaluation model for training to obtain a final evaluation model, inputting the segmentation feature map into the final evaluation model for evaluation to output a severe pancreatitis evaluation result; The denoising module 2 comprises: A block division submodule, used for dividing the abdominal CT image into blocks to obtain a plurality of image blocks; The clustering submodule is used to perform clustering processing on the image blocks using a K clustering algorithm to obtain a plurality of cluster clusters: The error submodule is used to construct the initial feature dictionary and set the initial sparse error, and determine the covariance matrix based on the initial feature field and the initial sparse error. : ; In the formula, represents the initial sparse error, represents the initial feature dictionary, express The initial diagonal matrix of order, The diagonal elements in are , ,… , Indicates Variation parameters; The iterative submodule is used to iteratively adjust the change parameters in the initial diagonal matrix and determine the CT processed image based on the adjusted change parameters.

[0054] The iterator module includes: An adjustment unit, configured to iteratively adjust the change parameters in the initial diagonal matrix, and iteratively update the sparse error based on the adjusted change parameters to obtain an iterative error : ; In the formula, , Respectively represent sequence The iterative error after iterations is represents the image blocks in the cluster, Indicates The feature dictionary after iterations, represents the number of column vectors in the feature dictionary, Indicates the change parameter adjustment The covariance matrix after the second step; The first iteration unit is used to Determine The updated sparse matrix after iterations : ; In the formula, Indicates Feature dictionary after iterations; The second iteration unit is used to The updated sparse matrix after iterations Determine the decomposition matrix : ; In the formula, represents the subscript matrix in the feature dictionary, Indicates The column vector in the feature dictionary after iterations; A decomposition unit, used to decompose the matrix Perform singular value decomposition to obtain a first orthogonal matrix and a second orthogonal matrix, and use the column vector in the first orthogonal matrix as the first The column vector of the feature dictionary after the iteration, the product of the first column vector of the second orthogonal matrix and the first eigenvalue is taken as the The column vector in the coefficient matrix after iterations; An output unit is used to repeat the iterative process until the iteration stop number is met and input the corresponding iterative feature dictionary and iterative coefficient matrix; Denoising unit, used for feature dictionary based on iteration And the coefficient matrix after iteration De-noising the image blocks in the cluster to obtain a CT processed image : ; In the formula, is Lagrangian sparse, represents the noisy area in the image block, represents the residual factor, represents the sparse identity matrix.

[0055] The fusion module 3 includes: A construction submodule is used to construct a first Gaussian filter and a plurality of second Gaussian filters in different directions, and input the first Gaussian filter and the second Gaussian filter into a deconvolution network for training and updating to obtain a first update filter and a second update filter; A first filtering submodule is used to input the nuclear magnetic resonance processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering, so as to obtain a first low-frequency nuclear magnetic resonance image and a plurality of second high-frequency nuclear magnetic resonance images; A second filtering submodule is used to input the CT processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering, so as to obtain a first low-frequency CT image and a plurality of second CT high-frequency images; A fusion submodule, configured to fuse the first low-frequency nuclear magnetic resonance image with the first low-frequency CT image to obtain a first fused image , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image ; A summation submodule is used to combine the first fused image input to the first update filter and the second fused image respectively input into the second update filter for convolution and summation to obtain a number of undetermined fused images; The frequency optimization submodule is used to iteratively determine the optimal frequency of iteration of the first update filter and the second update filter by using a particle swarm optimization algorithm, and use the undetermined fusion image corresponding to the optimal frequency of iteration as the fusion image.

[0056] The fusion submodule includes: A first weighting unit is used to calculate a first low-frequency fusion weight based on the first low-frequency nuclear magnetic resonance image and the first low-frequency CT image. Fusion weight with the second low frequency : ; ; In the formula, , Respectively represent the pixel coordinates of the first low-frequency MRI image and the first low-frequency CT image, , Represent the first and second adjustment parameters respectively, , Represent the first and second fixed parameters respectively, Respectively represent the gradient matrices in the first low-frequency MRI image and the first low-frequency CT image, represents the first Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; The first fusion unit is used to fuse the first low-frequency weight Fusion weight with the second low frequency The first low-frequency nuclear magnetic resonance image , the first low-frequency CT image Fusion is performed to obtain the first fused image : ; A filtering image unit is used to filter the image based on the second high frequency nuclear magnetic image. The second high-frequency CT image Determine the first high frequency filtered image With the second high frequency filtered image : ; ; In the formula, represents the characteristic scale, Indicates scale structural elements, , They represent morphological dilation and erosion operations respectively; A filter value unit, configured to filter the image based on the first high frequency With the second high frequency filtered image Confirm the first high frequency filter value respectively With the second high frequency filter value : ; ; In the formula, , Respectively represent the pixel coordinates of the first high-frequency filtered image and the second high-frequency filtered image, , Respectively represent the third and fourth adjustment parameters, , Respectively represent the third and fourth fixed parameters, Respectively represent the gradient matrices in the first high-frequency filtered image and the second high-frequency filtered image, represents the second Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; The second weighting unit is used to adjust the weight based on the first high frequency filter value. With the second high frequency filter value Calculate the first high frequency fusion weight Fusion weight with the second low frequency : ; ; The second fusion unit is used to fuse the first high frequency weight Fusion weight with the second low frequency The second high-frequency nuclear magnetic resonance image , Second high frequency CT image Fusion is performed to obtain a second fused image : .

[0057] The segmentation module 4 comprises: a division submodule, used for storing the pancreas template image into a template image set, and dividing the template image set into a first image set and a second image set according to a preset ratio; A first alignment submodule, used for arbitrarily selecting an image in the first image set as an initial pancreatic image, and aligning the remaining images in the first image set to the initial pancreatic image to obtain a plurality of registration alignment matrices, and calculating an average of the plurality of registration alignment matrices to obtain an average alignment matrix; A second alignment submodule is used to register and align all images in the second image set to the initial pancreas image, and select an image in the second image set that best matches the average alignment matrix as a standard image; The segmentation submodule is used to segment the fused image based on the standard image to obtain a segmentation feature map.

[0058] The segmentation submodule comprises: an alignment unit, configured to register and align the fused image to the standard image to obtain a fused alignment matrix, and calculate an average value of the fused alignment matrix and a plurality of the registration alignment matrices to obtain a fused average alignment matrix; A template output unit, which re-iterates the standard image and the registration alignment process based on the fused average alignment matrix until an iteration termination condition is met, and outputs a final template image; A binarization unit is used to extract the binarized image of the final template image to obtain a template image pair and calculate the likelihood image of the fused image. : ; In the formula, Represents the likelihood image The pixel value of a pixel, Indicates the fused image The pixel value of a pixel, represents the likelihood expectation, represents the likelihood variance; A segmentation unit is used for segmenting the image based on the likelihood Compared with the template image Determine segmentation feature map : ; In the formula, Represents and operations.

[0059] In some other embodiments of the present invention, the embodiments of the present invention provide the following technical solutions: a computer, comprising a memory 102, a processor 101, and a computer program stored in the memory 102 and executable on the processor 101, wherein the processor 101 implements the above-described image-based severe pancreatitis assessment method when executing the computer program.

[0060] Specifically, the processor 101 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.

[0061] Among them, the memory 102 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 102 may include a hard disk drive (HDD), a floppy disk drive, a solid state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 102 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 102 may be inside or outside the data processing device. In a specific embodiment, the memory 102 is a non-volatile memory. In a specific embodiment, the memory 102 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0062] The memory 102 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 101 .

[0063] The processor 101 implements the above-mentioned image-based severe pancreatitis assessment method by reading and executing computer program instructions stored in the memory 102 .

[0064] In some embodiments, the computer may further include a communication interface 103 and a bus 100. Figure 3 As shown, the processor 101, the memory 102, and the communication interface 103 are connected via a bus 100 and communicate with each other.

[0065] The communication interface 103 is used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present invention. The communication interface 103 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0066] The bus 100 includes hardware, software or both, and couples the components of the computer device to each other. The bus 100 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 100 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 100 may include one or more buses. Although embodiments of the present invention describe and illustrate a particular bus, the present invention contemplates any suitable bus or interconnect.

[0067] The computer can execute the image-based severe pancreatitis assessment method of the present invention based on the acquisition of the image-based severe pancreatitis assessment system, thereby realizing image-based severe pancreatitis assessment.

[0068] In some further embodiments of the present invention, in combination with the above-mentioned image-based severe pancreatitis assessment method, the embodiments of the present invention provide the following technical solutions: a storage medium having a computer program stored thereon, and the computer program implements the above-mentioned image-based severe pancreatitis assessment method when executed by a processor.

[0069] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0070] More specific examples of readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0071] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0072] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described 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.

[0073] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. An imaging-based method for assessing severe pancreatitis, characterized in that: include: Acquire an abdominal CT image and an abdominal MRI image of the patient, and sequentially perform image size adjustment, Gaussian filtering, and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; Performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; Acquire a pancreatic template image, and perform region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; A training image is obtained, and the training image is input into a preset evaluation model for training to obtain a final evaluation model, and the segmentation feature map is input into the final evaluation model for evaluation to output a severe pancreatitis evaluation result.

2. The imaging-based severe pancreatitis assessment method according to claim 1, characterized in that: The step of sequentially performing clustering processing and sparse denoising processing on the abdominal CT image to obtain a CT processed image comprises: Dividing the abdominal CT image into blocks to obtain a plurality of image blocks; The K clustering algorithm is used to cluster the image blocks to obtain a number of clusters: Construct the initial feature dictionary and set the initial sparse error, and determine the covariance matrix based on the initial feature field and the initial sparse error : ； In the formula, represents the initial sparse error, represents the initial feature dictionary, express The initial diagonal matrix of order, The diagonal elements in are , ,… , Indicates Variation parameters; The change parameters in the initial diagonal matrix are iteratively adjusted, and a CT processed image is determined based on the adjusted change parameters.

3. The imaging-based severe pancreatitis assessment method according to claim 2, characterized in that: The step of sequentially performing clustering processing and sparse denoising processing on the abdominal CT image to obtain a CT processed image comprises: Iteratively adjust the change parameters in the initial diagonal matrix, and iteratively update the sparse error based on the adjusted change parameters to obtain an iterative error : ； In the formula, , Respectively represent sequence The iterative error after iterations is represents the image blocks in the cluster, Indicates The feature dictionary after iterations, represents the number of column vectors in the feature dictionary, Indicates the change parameter adjustment The covariance matrix after the second step; Based on the iterative error Determine The updated sparse matrix after iterations : ； In the formula, Indicates Feature dictionary after iterations; Based on The updated sparse matrix after iterations Determine the decomposition matrix : ； In the formula, represents the subscript matrix in the feature dictionary, Indicates The column vector in the feature dictionary after iterations; For the decomposition matrix Perform singular value decomposition to obtain a first orthogonal matrix and a second orthogonal matrix, and use the column vector in the first orthogonal matrix as the first The column vector of the feature dictionary after the iteration, the product of the first column vector of the second orthogonal matrix and the first eigenvalue is taken as the The column vector in the coefficient matrix after iterations; Repeat the iterative process until the iteration stop number is met and input the corresponding iterative feature dictionary and iterative coefficient matrix; Based on the iterated feature dictionary And the coefficient matrix after iteration De-noising the image blocks in the cluster to obtain a CT processed image : ； In the formula, is Lagrangian sparse, represents the noisy area in the image block, represents the residual factor, represents the sparse identity matrix.

4. The imaging-based severe pancreatitis assessment method according to claim 1, characterized in that: The step of performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image comprises: Constructing a first Gaussian filter and a plurality of second Gaussian filters in different directions, and inputting the first Gaussian filter and the second Gaussian filter into a deconvolution network for training and updating to obtain a first update filter and a second update filter; Inputting the nuclear magnetic resonance processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering to obtain a first low-frequency nuclear magnetic resonance image and a plurality of second high-frequency nuclear magnetic resonance images; Inputting the CT processed image into a first Gaussian filter and a second Gaussian filter respectively for Gaussian filtering to obtain a first low-frequency CT image and a plurality of second CT high-frequency images; The first low-frequency nuclear magnetic resonance image is fused with the first low-frequency CT image to obtain a first fused image. , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image ; The first fused image input to the first update filter and the second fused image respectively input into the second update filter for convolution and summation to obtain a number of undetermined fused images; The particle swarm optimization algorithm is used to iteratively determine the optimal iteration frequency of the first update filter and the second update filter, and the undetermined fusion image corresponding to the optimal iteration frequency is used as the fusion image.

5. The imaging-based severe pancreatitis assessment method according to claim 1, characterized in that: The first low-frequency nuclear magnetic resonance image is fused with the first low-frequency CT image to obtain a first fused image. , the second high-frequency nuclear magnetic resonance image and the second high-frequency CT image are fused to obtain a second fused image The steps include: Calculate a first low-frequency fusion weight based on the first low-frequency nuclear magnetic resonance image and the first low-frequency CT image Fusion weight with the second low frequency : ； ； In the formula, , Respectively represent the pixel coordinates of the first low-frequency MRI image and the first low-frequency CT image, , Represent the first and second adjustment parameters respectively, , Represent the first and second fixed parameters respectively, Respectively represent the gradient matrices in the first low-frequency MRI image and the first low-frequency CT image, represents the first Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; Based on the first low frequency fusion weight Fusion weight with the second low frequency The first low-frequency nuclear magnetic resonance image , the first low-frequency CT image Fusion is performed to obtain the first fused image : ； Based on the second high-frequency nuclear magnetic image The second high-frequency CT image Determine the first high frequency filtered image With the second high frequency filtered image : ； ； In the formula, represents the characteristic scale, Indicates scale structural elements, , They represent morphological dilation and erosion operations respectively; Based on the first high frequency filtered image With the second high frequency filtered image Confirm the first high frequency filter value respectively With the second high frequency filter value : ； ； In the formula, , Respectively represent the pixel coordinates of the first high-frequency filtered image and the second high-frequency filtered image, , Respectively represent the third and fourth adjustment parameters, , Respectively represent the third and fourth fixed parameters, Respectively represent the gradient matrices in the first high-frequency filtered image and the second high-frequency filtered image, represents the second Gaussian filter, , Respectively The grayscale histogram of pixels at , , Respectively The grayscale histogram of pixels at , , Respectively expressed as The preset rectangular area is centered. , Respectively The pixel value at ; Based on the first high frequency filter value With the second high frequency filter value Calculate the first high frequency fusion weight Fusion weight with the second low frequency : ； ； Based on the first high frequency fusion weight Fusion weight with the second low frequency The second high-frequency nuclear magnetic resonance image , Second high frequency CT image Fusion is performed to obtain a second fused image : 。 6. The imaging-based severe pancreatitis assessment method according to claim 1, characterized in that: The step of performing region of interest segmentation on the fused image based on the pancreas template image to obtain a segmentation feature map comprises: Storing the pancreas template image in a template image set, and dividing the template image set into a first image set and a second image set according to a preset ratio; Randomly selecting an image in the first image set as an initial pancreatic image, and aligning the remaining images in the first image set to the initial pancreatic image to obtain a plurality of registration and alignment matrices, and calculating an average of the plurality of registration and alignment matrices to obtain an average alignment matrix; All images in the second image set are registered and aligned to the initial pancreas image, and the image in the second image set that best matches the average alignment matrix is ​​selected as the standard image; The fused image is segmented based on the standard image to obtain a segmentation feature map.

7. The imaging-based severe pancreatitis assessment method according to claim 1, characterized in that: The step of segmenting the fused image based on the standard image to obtain a segmentation feature map comprises: Aligning the fused image to the standard image to obtain a fused alignment matrix, and calculating an average value of the fused alignment matrix and a plurality of the registration alignment matrices to obtain a fused average alignment matrix; Re-iteratively determining the standard image and the registration alignment process based on the fused average alignment matrix until the iteration termination condition is met, and outputting the final template image; Extract the binary image of the final template image to obtain a template image pair, and calculate the likelihood image of the fused image : ； In the formula, Represents the likelihood image The pixel value of a pixel, Indicates the fused image The pixel value of a pixel, represents the likelihood expectation, represents the likelihood variance; Based on the likelihood image Compared with the template image Determine segmentation feature map : ； In the formula, Represents and operations.

8. An image-based severe pancreatitis assessment system, characterized in that: The system comprises: An enhancement module is used to obtain an abdominal CT image and an abdominal MRI image of a patient, and to sequentially perform image size adjustment, Gaussian filtering and image enhancement processing on the abdominal MRI image to obtain a MRI processed image; A denoising module, used for performing clustering processing and sparse denoising processing on the abdominal CT image in sequence to obtain a CT processed image; A fusion module, used for performing convolution fusion processing on the NMR processed image and the CT processed image to obtain a fused image; A segmentation module, used for acquiring a pancreatic template image, and performing region of interest segmentation on the fused image based on the pancreatic template image to obtain a segmentation feature map; The evaluation module is used to obtain training images, input the training images into a preset evaluation model for training to obtain a final evaluation model, input the segmentation feature map into the final evaluation model for evaluation to output a severe pancreatitis evaluation result.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the image-based severe pancreatitis assessment method according to any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the image-based severe pancreatitis assessment method according to any one of claims 1 to 7 is implemented.

Images