Parameter Determination Method, Device, Computer Equipment and Storage Medium

Abstract

The present application relates to a method, apparatus, computer device, and storage medium for parameter determination. The method includes: obtaining a first parameter distribution by using a pre-established normal tissue parameter distribution and target medical image data to be processed; processing the target medical image data by using a pre-trained region segmentation model to obtain target information corresponding to the target medical image data; the target information includes a target region where a lesion is located; determining a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table; and determining a target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution. By using this method, the difficulty of determining the initial value and boundary conditions of the parameters can be reduced, and the problem that the performance and accuracy of the optimization algorithm are affected due to poor determination of the initial value and boundary conditions can be avoided.

Description

Technical Field

[0001] This application relates to the technical field of parameter processing, and particularly to a method and apparatus for parameter determination, a computer device, and a storage medium. Background Art

[0002] MRI (Magnetic Resonance Imaging) examinations mainly based on ultra-fast imaging sequences can evaluate the functional status of organs and reveal physiological information within living organisms, collectively referred to as functional magnetic resonance imaging, or functional imaging techniques.

[0003] Currently, functional imaging techniques are widely used in clinical and scientific research. Among them, most functional imaging techniques require solving model parameters based on optimization algorithms to obtain corresponding parameter maps.

[0004] However, the solution of the optimization algorithm requires the initial values and boundary conditions of the parameters to be given. However, the initial values and boundary conditions of the parameters are not easy to determine, and if the initial values and boundary conditions are not determined well, it will also affect the performance and accuracy of the optimization algorithm. Summary of the Invention

[0005] Based on this, in order to solve the above problems, the embodiments of the present disclosure provide a method and apparatus for parameter determination, a computer device, and a storage medium, which can reduce the difficulty of determining the initial values and boundary conditions of the parameters and avoid the problem that the performance and accuracy of the optimization algorithm are affected due to the improper determination of the initial values and boundary conditions of the parameters.

[0006] In a first aspect, the embodiments of the present disclosure provide a method for parameter determination, the method comprising:

[0007] Using a pre-established normal tissue parameter distribution and target medical image data to be processed, obtain a first parameter distribution;

[0008] Using a pre-trained region segmentation model to process the target medical image data to obtain target information corresponding to the target medical image data; the target information includes a target region where a lesion is located;

[0009] Determine a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table;

[0010] According to the first parameter distribution and the second parameter distribution, determine a target parameter distribution corresponding to the target medical image data.

[0011] In one embodiment, the target information further includes the target lesion type and the target lesion degree corresponding to the target region. Determining the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table includes:

[0012] Searching for the second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0013] In one embodiment, before determining the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table, the method further includes:

[0014] Obtaining the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees:

[0015] Establishing an abnormal tissue parameter retrieval table according to the multiple abnormal tissue parameter distributions.

[0016] In one embodiment, processing the target medical image data by using the pre-trained region segmentation model to obtain the target information corresponding to the target medical image data includes:

[0017] Inputting the target medical image data into the region segmentation model to obtain the target region, the target lesion type and the target lesion degree output by the region segmentation model for region segmentation processing.

[0018] In one embodiment, the method further includes:

[0019] Obtaining a training sample set; the training sample set includes multiple sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include sample lesion masks, sample lesion types and sample lesion degrees;

[0020] Training a deep learning model based on the training sample set to obtain a region segmentation model.

[0021] In one embodiment, before using the pre-established normal tissue parameter distribution and the target medical image data to be processed, the method further includes:

[0022] Obtaining multiple first medical image data; the first medical image data is medical image data without lesions;

[0023] Performing spatial mapping processing on the multiple first medical image data to obtain multiple second medical image data in the standard space;

[0024] Performing parameter statistics according to the multiple second medical image data to obtain a normal tissue parameter distribution.

[0025] In one embodiment, determining the target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution includes:

[0026] Determine the parameter distribution of the non-target region in the target medical image data as the first parameter distribution, and determine the parameter distribution of the target region in the target medical image data as the second parameter distribution, so as to obtain the target parameter distribution corresponding to the target medical image data.

[0027] In one embodiment, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0028] In a second aspect, an embodiment of the present disclosure provides a parameter determination device, which includes:

[0029] A first parameter distribution determination module, configured to obtain a first parameter distribution by using a pre-established normal tissue parameter distribution and target medical image data to be processed;

[0030] A segmentation module, configured to process the target medical image data by using a pre-trained region segmentation model to obtain target information corresponding to the target medical image data; the target information includes the target region where the lesion is located;

[0031] A second parameter distribution determination module, configured to determine the second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table;

[0032] A target parameter distribution determination module, configured to determine the target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution.

[0033] In one embodiment, the target information further includes the target lesion type and the target lesion degree corresponding to the target region. The second parameter distribution determination module is specifically configured to find the second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0034] In one embodiment, the device further includes:

[0035] A parameter distribution acquisition module, configured to acquire abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees:

[0036] A retrieval table establishment module, configured to establish an abnormal tissue parameter retrieval table according to multiple abnormal tissue parameter distributions.

[0037] In one embodiment, the above-mentioned segmentation module is specifically configured to input the target medical image data into a region segmentation model, and obtain the target region, the target lesion type corresponding to the target region, and the target lesion degree output by the region segmentation model after performing region segmentation processing.

[0038] In one embodiment, the device further includes:

[0039] A sample set acquisition module, configured to acquire a training sample set; the training sample set includes a plurality of sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include a sample lesion mask, a sample lesion type, and a sample lesion degree;

[0040] A model training module, configured to train a deep learning model based on the training sample set to obtain a region segmentation model.

[0041] In one embodiment, the device further includes:

[0042] A data acquisition module, configured to acquire a plurality of first medical image data; the first medical image data is medical image data that does not contain lesions;

[0043] A mapping processing module, configured to perform spatial mapping processing on the plurality of first medical image data to obtain a plurality of second medical image data in a standard space;

[0044] A parameter statistics module, configured to perform parameter statistics according to the plurality of second medical image data to obtain a normal tissue parameter distribution.

[0045] In one embodiment, the above-mentioned target parameter distribution determination module is specifically configured to determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data, so as to obtain the target parameter distribution corresponding to the target medical image data.

[0046] In one embodiment, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0047] In a third aspect, an embodiment of the present disclosure provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect above is implemented.

[0048] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0049] The above-mentioned parameter determination method, device, computer device, and storage medium use a pre-established normal tissue parameter distribution and target medical image data to be processed to obtain a first parameter distribution; use a pre-trained region segmentation model to process the target medical image data to obtain target information corresponding to the target medical image data; determine a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table; and determine a target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution. In the embodiments of the present disclosure, the target parameter distribution is determined by separately determining the first parameter distribution of normal tissues and the second parameter distribution of abnormal tissues. According to the target parameter distribution, the initial values and boundary conditions of the parameters of the optimization algorithm can be more easily determined, thereby reducing the difficulty of determining the initial values and boundary conditions of the parameters. Moreover, since the target parameter distribution includes not only the parameter distribution of normal tissues but also the parameter distribution of abnormal tissues, the accuracy of the target parameter distribution is relatively high, and problems that the performance and accuracy of the optimization algorithm are affected due to improper determination of the initial values and boundary conditions of the parameters can be avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 FIG. is an application environment diagram of the parameter determination method in an embodiment;

[0051] Figure 2 FIG. is a flowchart of the parameter determination method in an embodiment;

[0052] Figure 3a FIG. is a schematic diagram of the image data of functional imaging in an embodiment;

[0053] Figure 3b FIG. is a schematic diagram of the signal distribution curve in an embodiment;

[0054] Figure 3c FIG. is a schematic diagram of the first parameter distribution in an embodiment;

[0055] Figure 3d FIG. is a schematic diagram of the target region where the lesion is located in an embodiment;

[0056] Figure 3e FIG. is a schematic diagram of the abnormal tissue parameter retrieval table in an embodiment;

[0057] Figure 3f FIG. is a schematic diagram of the target parameter distribution in an embodiment;

[0058] Figure 4 FIG. is a flowchart of the steps of establishing an abnormal tissue parameter retrieval table in an embodiment;

[0059] Figure 5 FIG. is a flowchart of the model training steps in an embodiment;

[0060] Figure 6 It is a schematic flow chart of the steps for determining the normal tissue parameter distribution in an embodiment;

[0061] Figure 7 It is a schematic diagram of medical image data in an embodiment;

[0062] Figure 8 It is a schematic flow chart of the parameter determination method in another embodiment;

[0063] Figure 9 It is a structural block diagram of the parameter determination device in an embodiment;

[0064] Figure 10 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0065] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] The parameter determination method provided by the present application can be applied to, for example, Figure 1 the application environment shown in the figure. This application environment may include a terminal 101 and multiple medical scanning devices 102. Among them, the terminal 101 can communicate with the medical scanning device 102 through a network. The above terminal 101 can be, but is not limited to, various personal computers, laptop computers and tablet computers, and the above medical scanning device 102 can be, but is not limited to, CT (Computed Tomography) devices, PET (Positron Emission Computed Tomography)-CT devices and MR (Magnetic Resonance) devices.

[0067] This application environment may further include a PACS (Picture Archiving and Communication Systems) server 103, and both the terminal 101 and the medical scanning device 102 can communicate with the PACS server 103 through a network. The above PACS server 103 can be implemented by an independent server or a server cluster composed of multiple servers.

[0068] In one embodiment, as shown in Figure 2 the figure, a parameter determination method is provided. Taking the method applied to the Figure 1 terminal in the figure as an example, the method includes the following steps:

[0069] Step 201: Using the pre-established normal tissue parameter distribution and the target medical image data to be processed, obtain the first parameter distribution.

[0070] Among them, the parameter distribution is the statistical distribution of a certain parameter of the pixel points in the image data, including at least one of the parameter mean, parameter variance, parameter maximum value, and parameter minimum value. The parameter can include Apparent Diffusion Coefficient (ADC), etc. The target medical image data to be processed can be the original CT image data, MR image data, or the image data of functional imaging.

[0071] The acquisition process of the target medical image data can include: performing multiple scans of the target scanning part with different parameters to obtain multiple image data of functional imaging, as Figure 3a shown. In the image data of functional imaging, the same tissue corresponds to different feedback signals. Therefore, the signal distribution curve can be determined according to the scanning parameters and the feedback signals, as Figure 3b shown.

[0072] Exemplarily, the signal distribution curve can be IVIM (Intro-Voxel Incoherent Movement). IVIM is a distribution curve related to the main magnetic field of the magnetic resonance device, and the IVIM model is S b = S0 * ((1 - f) * e -b*D + f * e -b*Dstar ), and the feedback signal value is S0 when b = 0. Among them, b is the diffusion weighting factor, S b is the feedback signal value with the diffusion weighting factor b, S0 is the feedback signal value without the diffusion weighting factor, D is the true diffusion, reflecting the simple molecular diffusion, f is the perfusion fraction related to the microcirculation, and D* is the pseudo-diffusion factor related to the perfusion.

[0073] The terminal can pre-establish the normal tissue parameter distribution. Then, perform registration processing on the normal tissue parameter distribution and the target medical image data to be processed to obtain the first parameter distribution corresponding to the target medical image data, as Figure 3c shown.

[0074] The above registration processing can include at least one of rigid registration processing and elastic registration processing, and the embodiments of the present disclosure do not limit this.

[0075] Step 202: Using the pre-trained region segmentation model to process the target medical image data to obtain the target information corresponding to the target medical image data.

[0076] Among them, the target information includes the target region where the lesion is located.

[0077] A pre-trained region segmentation model can be set in the terminal. After obtaining the target medical image data, the region segmentation model is used to perform region segmentation processing on the target medical image data to obtain the target region where the lesion is located in the target medical image data, as Figure 3d shown.

[0078] The present disclosure embodiment does not limit the structure and type of the region segmentation model, and can be set according to actual situations.

[0079] Step 203: Determine the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table.

[0080] The terminal pre-establishes an abnormal tissue parameter retrieval table. After determining the target region where the lesion is located, the second parameter distribution corresponding to the target region can be found in the abnormal tissue parameter retrieval table according to the lesion type, lesion degree, etc. of the lesion.

[0081] Among them, the abnormal tissue parameter retrieval table is as Figure 3e shown.

[0082] Step 204: Determine the target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution.

[0083] Determine the parameter distribution of the normal region outside the lesion in the target medical image data according to the first parameter distribution, and determine the parameter distribution of the target region where the lesion is located according to the second parameter distribution; then, determine the target parameter distribution corresponding to the target medical image data according to the parameter distribution of the normal region and the parameter distribution of the target region, as Figure 3f shown.

[0084] It can be understood that after determining the target parameter distribution, it is relatively easy to determine the initial values and boundary conditions of the parameters of the optimization algorithm, that is, the difficulty of determining the initial values and boundary conditions of the parameters is reduced. Moreover, since the target parameter distribution is determined according to the empirical distribution of normal tissues and the abnormal tissue parameter retrieval table, the problem that the performance and accuracy of the optimization algorithm are affected due to improper determination of the initial values and boundary conditions of the parameters can be avoided.

[0085] The above optimization algorithm may include the LM (Levenberg-Marquardt) algorithm, the TRF (Trust Region Reflective) algorithm, and the interior point method. Among them, the unconstrained problem can use the LM algorithm, the boundary constraint problem can use the TRF algorithm (as shown in the figure, the boundary constraint is x>3, y<2), and the inequality constraint can use the interior point method (such as x + y>0). The present disclosure embodiment does not limit the optimization algorithm.

[0086] In the above parameter determination method, a first parameter distribution is obtained by using a pre-established normal tissue parameter distribution and the target medical image data to be processed; the target medical image data is processed by using a pre-trained region segmentation model to obtain target information corresponding to the target medical image data; a second parameter distribution corresponding to the target region is determined in a pre-established abnormal tissue parameter retrieval table; and a target parameter distribution corresponding to the target medical image data is determined according to the first parameter distribution and the second parameter distribution. In the embodiment of the present disclosure, the target parameter distribution is determined by separately determining the first parameter distribution of the normal tissue and the second parameter distribution of the abnormal tissue. According to the target parameter distribution, the initial value and boundary conditions of the optimization algorithm can be relatively easily determined, thereby reducing the difficulty of determining the initial value and boundary conditions of the parameters. Moreover, since the target parameter distribution includes not only the parameter distribution of the normal tissue but also the parameter distribution of the abnormal tissue, the accuracy of the target parameter distribution is relatively high, and the problem that the performance and accuracy of the optimization algorithm are affected due to improper determination of the initial value and boundary conditions of the parameters can be avoided.

[0087] In one embodiment, the target information further includes the target lesion type and the target lesion degree corresponding to the target region. The process of determining the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table may include: searching for the second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0088] When the terminal uses the region segmentation model to perform region segmentation on the target medical image data, it can not only obtain the target region where the lesion is located, but also obtain the target lesion type and the target lesion degree corresponding to the target region. For example Figure 3d as shown, in addition to determining the target region where the lesion is located in the figure, it is also determined that the target lesion type corresponding to the target region is glioblastoma, and the target lesion degree is grade three.

[0089] In the abnormal tissue parameter retrieval table as Figure 3e shown, the terminal can search for the target lesion type among multiple lesion types, and then search for the target lesion degree among multiple lesion degrees corresponding to the target lesion type; then, the parameter distribution corresponding to the target lesion type and the target lesion degree is determined as the second parameter distribution corresponding to the target region.

[0090] In one of the embodiments, as Figure 3e shown, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0091] In the above embodiments, the second parameter distribution corresponding to the target region is found in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree. Since the abnormal tissue parameter retrieval table is established in advance, it is relatively easy to determine the second parameter distribution corresponding to the target region where the lesion is located in the target medical image data, and the second parameter distribution is also relatively accurate. In this way, the determination efficiency and accuracy of the target parameter distribution can be improved, thereby reducing the difficulty of determining the initial parameter value and the boundary condition, and avoiding the problem that the performance and accuracy of the optimization algorithm are affected due to the improper determination of the initial parameter value and the boundary condition.

[0092] In one embodiment, as Figure 4 shown, before the step of determining the second parameter distribution corresponding to the target region in the above-mentioned pre-established abnormal tissue parameter retrieval table, the embodiments of the present disclosure may further include the following steps:

[0093] Step 301, obtain the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees.

[0094] The terminal can obtain multiple medical image data from the medical image database, and then obtain the abnormal tissue parameter distributions corresponding to multiple lesion degrees in each lesion type. Alternatively, the terminal obtains relevant materials and documents of the medical image data, and then extracts the abnormal tissue parameter distributions corresponding to multiple lesion degrees in each lesion type from the relevant materials and documents. The embodiments of the present disclosure do not limit the acquisition method, and can be selected according to the actual situation.

[0095] Exemplarily, for glioblastoma, the terminal obtains the abnormal tissue parameter distribution corresponding to grade I glioblastoma, the abnormal tissue parameter distribution corresponding to grade II glioblastoma, and the abnormal tissue parameter distribution corresponding to grade III glioblastoma. And so on, the terminal can obtain the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees.

[0096] Step 302, establish an abnormal tissue parameter retrieval table according to the multiple abnormal tissue parameter distributions.

[0097] The terminal can establish the retrieval columns of the abnormal tissue parameter retrieval table according to the lesion type and the lesion degree, and then fill the abnormal tissue parameter distributions into the corresponding retrieval columns, so that in practical applications, the second parameter distribution corresponding to the target region where the lesion is located can be retrieved according to the target lesion type and the target lesion degree.

[0098] In the above embodiments, the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees are obtained; according to the multiple abnormal tissue parameter distributions, an abnormal tissue parameter retrieval table is established. By pre-establishing the abnormal tissue parameter retrieval table in the embodiments of the present disclosure, the determination efficiency and accuracy of the abnormal tissue parameter distribution are improved, thereby improving the determination efficiency and accuracy of the target parameter distribution, reducing the difficulty of determining the initial parameter value and boundary conditions, and avoiding the problem that the performance and accuracy of the optimization algorithm are affected by the improper determination of the initial parameter value and boundary conditions.

[0099] In one embodiment, the process of using the pre-trained region segmentation model to process the target medical image data to obtain the target information corresponding to the target medical image data may include: inputting the target medical image data into the region segmentation model to obtain the target region output by the region segmentation model for region segmentation processing, the target lesion type corresponding to the target region, and the target lesion degree.

[0100] Exemplarily, when the target medical image data is input into the region segmentation model, the region segmentation model performs region segmentation processing on the target medical image data and outputs the Figure 3d target region as shown, and the target region corresponds to grade III glioma.

[0101] As Figure 5 shown, before using the region segmentation model for region segmentation processing, the embodiments of the present disclosure may further include a model training process, such as the following steps:

[0102] Step 401, obtain a training sample set.

[0103] Among them, the training sample set includes multiple sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include sample lesion masks, sample lesion types, and sample lesion degrees.

[0104] The terminal can obtain multiple medical image data containing lesions from a medical scanning device or a PACS server, and use the medical image data containing lesions as the sample image data. Then, for any sample image data, the terminal can obtain the sample lesion contour manually drawn by the user in the sample image data, as well as the sample lesion type and sample lesion degree input by the user. Then, the sample lesion mask, sample lesion type, and sample lesion degree are used as the annotations corresponding to the sample image data.

[0105] In practical applications, other methods may also be adopted to obtain the sample image data and the annotations corresponding to the sample image data, and the embodiments of the present disclosure do not limit this.

[0106] Step 402, train a deep learning model based on the training sample set to obtain a region segmentation model.

[0107] The terminal inputs a sample image data into a deep learning model to obtain a training result output by the deep learning model; calculates a loss value between the training result and the annotation corresponding to the sample image data by using a preset loss function. If the loss value does not meet the preset convergence condition, the adjustable parameters in the deep learning model are adjusted, and another sample image data is input into the deep learning model for continuous training. The training is ended until the loss value between the training result output by the deep learning model and the annotation meets the preset convergence condition, and the deep learning model at the end of the training is determined as the region segmentation model. The embodiments of the present disclosure do not limit the loss function and the preset convergence condition, and can be set according to actual situations.

[0108] In the above embodiments, before performing region segmentation processing by using the region segmentation model, a training sample set is obtained; the deep learning model is trained based on the training sample set to obtain the region segmentation model. Then, the target medical image data is input into the region segmentation model to obtain the target region output by the region segmentation model for region segmentation processing, the target lesion type corresponding to the target region, and the target lesion degree. The embodiments of the present disclosure can relatively easily identify the target region where the abnormal tissue is located from the target medical image data by pre-training the region segmentation model, so as to determine the parameter distribution of the abnormal tissue. Since the parameter distributions of the normal tissue and the abnormal tissue are determined separately, the finally obtained target parameter distribution is relatively accurate, and it will be more appropriate to determine the parameter initial values and boundary conditions of the optimization algorithm according to the target parameter distribution, avoiding the problem that the performance and accuracy of the optimization algorithm are affected due to the improper determination of the parameter initial values and boundary conditions.

[0109] In one embodiment, as Figure 6 shown, before the above-mentioned use of the pre-established normal tissue parameter distribution and the target medical image data to be processed, the embodiments of the present disclosure may further include:

[0110] Step 501, obtaining a plurality of first medical image data.

[0111] Among them, the first medical image data is medical image data without lesions. It can be understood that in the first medical image data, all tissues are normal tissues.

[0112] The terminal can obtain a plurality of first medical image data from a medical scanning device or a PACS server. The embodiments of the present disclosure do not limit the obtaining method.

[0113] Step 502, performing spatial mapping processing on the plurality of first medical image data to obtain a plurality of second medical image data in the standard space.

[0114] Since the spatial coordinates corresponding to multiple first medical image data may be different, which is not conducive to parameter statistics, therefore, spatial mapping processing is performed on multiple first medical image data, and each first medical image data is mapped to the standard space to obtain multiple second medical image data, as Figure 7 shown.

[0115] Exemplarily, each first medical image data is mapped to the brain MNI (Montreal Neurological Institute) space to obtain multiple second medical image data. It can be understood that the standard space is determined according to the actual situation, not limited to the brain MNI space.

[0116] Step 503, perform parameter statistics according to multiple second medical image data to obtain the normal tissue parameter distribution.

[0117] Among them, the parameter distribution includes at least one of parameter mean, parameter variance, parameter maximum value, and parameter minimum value.

[0118] The terminal can calculate the parameter mean, parameter variance, parameter maximum value, and parameter minimum value for each tissue or each position in the second medical image data, and obtain the parameter mean, parameter variance, parameter maximum value, and parameter minimum value of each tissue or each position, so as to determine the normal tissue parameter distribution.

[0119] In the above embodiment, multiple first medical image data are obtained; spatial mapping processing is performed on multiple first medical image data to obtain multiple second medical image data in the standard space; parameter statistics are performed according to multiple second medical image data to obtain the normal tissue parameter distribution. The embodiment of the present disclosure obtains the normal tissue parameter distribution by statistically analyzing the parameters of the normal tissue. Subsequently, using the normal tissue parameter distribution to determine the target parameter distribution can improve the determination efficiency and accuracy of the target parameter distribution. Moreover, determining the parameter initial values and boundary conditions of the optimization algorithm according to the target parameter distribution can also avoid the problem that the performance and accuracy of the optimization algorithm are affected due to improper determination of the parameter initial values and boundary conditions.

[0120] In one embodiment, as Figure 8 shown, a parameter determination method is provided. Taking the terminal in Figure 1 as an example for illustration, it includes the following steps:

[0121] Step 601, obtain multiple first medical image data.

[0122] Among them, the first medical image data is medical image data that does not contain lesions.

[0123] Step 602: Perform spatial mapping processing on multiple pieces of first medical image data to obtain multiple pieces of second medical image data in the standard space.

[0124] Step 603: Conduct parameter statistics based on multiple pieces of second medical image data to obtain the normal tissue parameter distribution.

[0125] Step 604: Utilize the pre-established normal tissue parameter distribution and the target medical image data to be processed to obtain the first parameter distribution.

[0126] Step 605: Input the target medical image data into the region segmentation model to obtain the target region output by the region segmentation model for region segmentation processing, the corresponding target lesion type of the target region, and the target lesion degree.

[0127] Step 606: Obtain the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees.

[0128] Step 607: Establish an abnormal tissue parameter retrieval table based on multiple abnormal tissue parameter distributions.

[0129] Step 608: Determine the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table.

[0130] Step 609: Determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data to obtain the target parameter distribution corresponding to the target medical image data.

[0131] In the above embodiments, the normal tissue parameter distribution and the abnormal tissue parameter retrieval table are obtained in advance; when determining the parameter distribution of the target medical image data, first use the region segmentation model to identify the region where the abnormal tissue is located, that is, the region where the lesion is located, from the target medical image data; then, determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data to obtain the target parameter distribution corresponding to the target medical image data. On the one hand, using the pre-obtained normal tissue parameter distribution and abnormal tissue parameter retrieval table to determine the target parameter distribution can improve the determination efficiency of the target parameter distribution; on the other hand, separately determining the parameter distributions of normal tissue and abnormal tissue can improve the accuracy of the target parameter distribution. In this way, when subsequently determining the initial values and boundary conditions of the parameters of the optimization algorithm based on the target parameter distribution, not only can the difficulty of the initial values and boundary conditions of the parameters be reduced, but also relatively appropriate initial values and boundary conditions of the parameters can be determined, thereby avoiding the problem that the performance and accuracy of the optimization algorithm are affected due to the improper determination of the initial values and boundary conditions of the parameters.

[0132] It should be understood that althoughFigures 2 to 8 The steps in the flowchart are shown in sequence according to the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figures 2 to 8 at least a part of the steps in Figures 2 to 8 may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0133] In one embodiment, as Figure 9 shown, a parameter determination device is provided, including:

[0134] A first parameter distribution determination module 701, configured to obtain a first parameter distribution by using a pre-established normal tissue parameter distribution and target medical image data to be processed;

[0135] A segmentation module 702, configured to process the target medical image data by using a pre-trained region segmentation model to obtain target information corresponding to the target medical image data; the target information includes a target region where a lesion is located;

[0136] A second parameter distribution determination module 703, configured to determine a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table;

[0137] A target parameter distribution determination module 704, configured to determine a target parameter distribution corresponding to the target medical image data according to the first parameter distribution and the second parameter distribution.

[0138] In one of the embodiments, the target information further includes a target lesion type and a target lesion degree corresponding to the target region. The above-mentioned second parameter distribution determination module 703 is specifically configured to find out the second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0139] In one of the embodiments, the device further includes:

[0140] A parameter distribution acquisition module, configured to acquire abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees:

[0141] A retrieval table establishment module, configured to establish an abnormal tissue parameter retrieval table according to the multiple abnormal tissue parameter distributions.

[0142] In one embodiment, the above-mentioned segmentation module 702 is specifically configured to input the target medical image data into a region segmentation model, and obtain the target region, the target lesion type corresponding to the target region, and the target lesion degree output by the region segmentation model after performing region segmentation processing.

[0143] In one embodiment, the device further includes:

[0144] A sample set acquisition module, configured to acquire a training sample set; the training sample set includes a plurality of sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include sample lesion masks, sample lesion types, and sample lesion degrees;

[0145] A model training module, configured to train a deep learning model based on the training sample set to obtain a region segmentation model.

[0146] In one embodiment, the device further includes:

[0147] A data acquisition module, configured to acquire a plurality of first medical image data; the first medical image data is medical image data that does not contain lesions;

[0148] A mapping processing module, configured to perform spatial mapping processing on the plurality of first medical image data to obtain a plurality of second medical image data in a standard space;

[0149] A parameter statistics module, configured to perform parameter statistics based on the plurality of second medical image data to obtain a normal tissue parameter distribution.

[0150] In one embodiment, the above-mentioned target parameter distribution determination module 704 is specifically configured to determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data, so as to obtain the target parameter distribution corresponding to the target medical image data.

[0151] In one embodiment, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0152] For the specific limitations on the parameter determination device, reference may be made to the limitations on the parameter determination method in the foregoing text, which will not be elaborated here. Each module in the above-mentioned parameter determination device can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0153] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structural diagram may be as shown in Figure 10 . The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be implemented through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a parameter determination method. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device may be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0154] Those skilled in the art can understand that Figure 10 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0155] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0156] Using a pre-established normal tissue parameter distribution and target medical image data to be processed, a first parameter distribution is obtained;

[0157] Using a pre-trained region segmentation model to process the target medical image data to obtain target information corresponding to the target medical image data; the target information includes the target region where the lesion is located;

[0158] Determining a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table;

[0159] According to the first parameter distribution and the second parameter distribution, determining a target parameter distribution corresponding to the target medical image data.

[0160] In one embodiment, the target information further includes the target lesion type and the target lesion degree corresponding to the target region. When the processor executes the computer program, the following steps are further implemented:

[0161] Search for the second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0162] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0163] Obtain the abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees:

[0164] Establish an abnormal tissue parameter retrieval table according to the multiple abnormal tissue parameter distributions.

[0165] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0166] Input the target medical image data into the region segmentation model to obtain the target region, the target lesion type and the target lesion degree corresponding to the target region output by the region segmentation model for region segmentation processing.

[0167] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0168] Obtain a training sample set; the training sample set includes multiple sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include sample lesion masks, sample lesion types and sample lesion degrees;

[0169] Train a deep learning model based on the training sample set to obtain a region segmentation model.

[0170] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0171] Obtain multiple first medical image data; the first medical image data is medical image data without lesions;

[0172] Perform spatial mapping processing on the multiple first medical image data to obtain multiple second medical image data in the standard space;

[0173] Perform parameter statistics according to the multiple second medical image data to obtain the normal tissue parameter distribution.

[0174] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0175] Determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data to obtain the target parameter distribution corresponding to the target medical image data.

[0176] In one embodiment, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0177] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0178] Using a pre-established normal tissue parameter distribution and target medical image data to be processed, a first parameter distribution is obtained;

[0179] Using a pre-trained region segmentation model to process the target medical image data, target information corresponding to the target medical image data is obtained; the target information includes a target region where a lesion is located;

[0180] Determining a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table;

[0181] According to the first parameter distribution and the second parameter distribution, a target parameter distribution corresponding to the target medical image data is determined.

[0182] In one embodiment, the target information further includes a target lesion type and a target lesion degree corresponding to the target region. When the computer program is executed by a processor, the following steps are further implemented:

[0183] Searching for a second parameter distribution corresponding to the target region in the abnormal tissue parameter retrieval table according to the target lesion type and the target lesion degree.

[0184] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0185] Obtaining abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees:

[0186] According to the multiple abnormal tissue parameter distributions, an abnormal tissue parameter retrieval table is established.

[0187] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0188] Inputting the target medical image data into the region segmentation model to obtain a target region, a target lesion type, and a target lesion degree corresponding to the target region output by the region segmentation model for region segmentation processing.

[0189] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0190] Obtain a training sample set; the training sample set includes a plurality of sample image data and the annotations corresponding to each sample image data; the sample image data is medical image data containing lesions, and the annotations include sample lesion masks, sample lesion types, and sample lesion degrees;

[0191] Train a deep learning model based on the training sample set to obtain a region segmentation model.

[0192] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0193] Obtain a plurality of first medical image data; the first medical image data is medical image data that does not contain lesions;

[0194] Perform spatial mapping processing on the plurality of first medical image data to obtain a plurality of second medical image data in the standard space;

[0195] Perform parameter statistics based on the plurality of second medical image data to obtain the parameter distribution of normal tissues.

[0196] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0197] Determine the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data, to obtain the target parameter distribution corresponding to the target medical image data.

[0198] In one embodiment, the parameter distribution includes at least one of a parameter mean, a parameter variance, a parameter maximum value, and a parameter minimum value.

[0199] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0200] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0201] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for determining parameters, characterized in that, The method includes: Using a pre-established normal tissue parameter distribution and the target medical image data to be processed to obtain a first parameter distribution, where the first parameter distribution is obtained through registration processing based on the normal tissue parameter distribution and the target medical image data to be processed; Using a pre-trained region segmentation model to process the target medical image data to obtain target information corresponding to the target medical image data; the target information includes the target region where the lesion is located; Determining a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table; Determining the first parameter distribution as the parameter distribution of the non-target region in the target medical image data, and determining the second parameter distribution as the parameter distribution of the target region in the target medical image data, to obtain a target parameter distribution corresponding to the target medical image data.

2. The method according to claim 1, wherein The target information further includes the target lesion type and the target lesion degree corresponding to the target region, and determining the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table includes: Searching in the abnormal tissue parameter retrieval table for the second parameter distribution corresponding to the target region according to the target lesion type and the target lesion degree.

3. The method according to claim 2, wherein Before determining the second parameter distribution corresponding to the target region in the pre-established abnormal tissue parameter retrieval table, the method further includes: Obtaining abnormal tissue parameter distributions corresponding to multiple lesion types and multiple lesion degrees: Establishing the abnormal tissue parameter retrieval table according to the multiple abnormal tissue parameter distributions.

4. The method according to claim 2, wherein Using the pre-trained region segmentation model to process the target medical image data to obtain the target information corresponding to the target medical image data includes: Inputting the target medical image data into the region segmentation model to obtain the target region, the target lesion type and the target lesion degree output by the region segmentation model for region segmentation processing.

5. The method according to claim 1, characterized in that Before using the pre-established normal tissue parameter distribution and the target medical image data to be processed, the method further includes: Obtaining multiple first medical image data; the first medical image data are medical image data without lesions; Performing spatial mapping processing on the multiple first medical image data to obtain multiple second medical image data in the standard space; Performing parameter statistics according to the multiple second medical image data to obtain the normal tissue parameter distribution.

6. The method according to any one of claims 1-5, characterized in that, The target parameter distribution is used to determine the parameter initial value and boundary conditions of the optimization algorithm.

7. The method according to claim 1, characterized in that The parameter distribution includes at least one of parameter mean, parameter variance, parameter maximum value, and parameter minimum value.

8. A parameter determination device, characterized in that, The device includes: A first parameter distribution determination module, configured to use a pre-established normal tissue parameter distribution and the target medical image data to be processed to obtain a first parameter distribution, where the first parameter distribution is obtained through registration processing based on the normal tissue parameter distribution and the target medical image data to be processed; A segmentation module, configured to process the target medical image data by using a pre-trained region segmentation model, so as to obtain target information corresponding to the target medical image data; the target information includes a target region where a lesion is located; A second parameter distribution determination module, configured to determine a second parameter distribution corresponding to the target region in a pre-established abnormal tissue parameter retrieval table; A target parameter distribution determination module, configured to determine the first parameter distribution as the parameter distribution of a non-target region in the target medical image data, and determine the second parameter distribution as the parameter distribution of the target region in the target medical image data, so as to obtain a target parameter distribution corresponding to the target medical image data.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images