Medical assistance information generation method, device, computer device and storage medium

Abstract

The present invention relates to the field of image classification, and discloses a method, device, computer device and storage medium for generating medical auxiliary information based on image parsing. The method includes: obtaining a medical image to be parsed; processing the medical image through an image parsing model to generate pathological analysis information, where the pathological analysis information includes a number of partition parsing information, the medical image is divided into a number of pathological partitions, and one pathological partition corresponds to one partition parsing information; performing semantic parsing on the pathological analysis information through a pathological analysis model to generate disease condition reference information; obtaining a reference image corresponding to the disease condition reference information; comparing the differences between the medical image and the reference image to generate a comparison result; and importing the pathological analysis information, disease condition reference information and comparison result into a reference information template to generate auxiliary reference information for the medical image. The present invention improves the recognition ability of medical images and can provide more accurate reference opinions for medical personnel.

Description

Technical Field

[0001] The present invention relates to the field of image classification, and particularly to a method, device, computer device and storage medium for generating medical auxiliary information based on image analysis. Background Art

[0002] At present, the reading of medical images in the medical field generally relies on the personal experience of doctors for judgment, or the machine scans to identify fixed-site shadows (lesions), and no specific doctor suggestions or evaluations are provided. Although in the prior art, medical images can match some conventional lesion description information, the matched lesion description information cannot accurately describe the disease characteristics of the medical images. This will cause certain misguidance to medical staff. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device and storage medium for generating medical auxiliary information based on image analysis, so as to improve the recognition ability of medical images and provide more accurate reference opinions for medical staff.

[0004] A method for generating medical auxiliary information based on image analysis includes:

[0005] Obtaining a medical image to be analyzed;

[0006] Processing the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one pathological partition corresponds to one partition analysis information;

[0007] Performing semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information;

[0008] Obtaining a reference image corresponding to the disease condition reference information;

[0009] Comparing the differences between the medical image and the reference image to generate a comparison result;

[0010] Importing the pathological analysis information, the disease condition reference information and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0011] A device for generating medical auxiliary information based on image analysis includes:

[0012] An obtaining module, configured to obtain a medical image to be analyzed;

[0013] An image analysis module, configured to process the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes a number of partition analysis information, the medical image is divided into a number of pathological partitions, and one pathological partition corresponds to one piece of partition analysis information;

[0014] A semantic analysis module, configured to perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information;

[0015] A reference image acquisition module, configured to acquire a reference image corresponding to the disease condition reference information;

[0016] An image comparison module, configured to compare the differences between the medical image and the reference image to generate a comparison result;

[0017] A reference information generation module, configured to import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0018] A computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, where when the processor executes the computer-readable instructions, the above-mentioned method for generating medical auxiliary information based on image analysis is implemented.

[0019] One or more readable storage media storing computer-readable instructions, where when the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the above-mentioned method for generating medical auxiliary information based on image analysis.

[0020] The above-mentioned method, device, computer equipment and storage medium for generating medical assistance information based on image analysis obtain a medical image to be analyzed, and identify corresponding pathological analysis information through the medical image. The medical image is processed by an image analysis model to generate pathological analysis information including several pathological regions. Here, by analyzing the medical image in the form of pathological regions, fine local pathological information can be obtained. The pathological analysis information is semantically analyzed by a pathological analysis model to generate disease condition reference information. Here, by using semantic recognition to interpret the pathological analysis information, more accurate disease condition reference information can be obtained. A reference image corresponding to the disease condition reference information is obtained. Here, the reference image can reflect the typical characteristics of a certain disease type. The difference between the medical image and the reference image is compared to generate a comparison result. Here, the comparison result can reflect the degree of approximation of the two images. The pathological analysis information, the disease condition reference information and the comparison result are imported into a reference information template to generate the auxiliary reference information of the medical image. Here, the generated auxiliary reference information is richer and more accurate. The present invention uses two neural network models to analyze the medical image in stages and uses a reference image to verify the medical image, ensuring the accuracy of the pathological analysis information and the disease condition reference information. The present invention improves the recognition ability of medical images and can provide more accurate reference opinions for medical personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 is a schematic diagram of an application environment of a method for generating medical assistance information based on image analysis according to an embodiment of the present invention;

[0023] Figure 2 is a schematic flowchart of a method for generating medical assistance information based on image analysis according to an embodiment of the present invention;

[0024] Figure 3 is a schematic structural diagram of a device for generating medical assistance information based on image analysis according to an embodiment of the present invention;

[0025] Figure 4 is a schematic diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0027] The medical auxiliary information generation method based on image analysis provided in this embodiment can be applied in an application environment such as Figure 1 where the client communicates with the server. Among them, the client includes but is not limited to various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented by an independent server or a server cluster composed of multiple servers.

[0028] In one embodiment, as Figure 2 shown, a medical auxiliary information generation method based on image analysis is provided. Taking the example of the method applied in the Figure 1 server, the following steps S10 - S60 are included.

[0029] S10. Obtain the medical image to be analyzed.

[0030] Understandably, the medical image can be a medical picture, such as a CT (Computed Tomography) image, an X - ray film, etc., or a medical video, such as a gastroscopy video, etc. The medical image to be analyzed here can be a pre - processed image file. For example, the medical picture to be analyzed can be processed into a picture of a preset specification size; the medical video to be analyzed can be processed into an image sequence including multiple video frames.

[0031] S20. Process the medical image through an image analysis model to generate pathological analysis information. The pathological analysis information contains several partition analysis information. The medical image is divided into several pathological partitions, and one of the pathological partitions corresponds to one of the partition analysis information.

[0032] Understandably, the image analysis model can be a neural network model trained based on a large number of historical medical images. When the image analysis model analyzes a medical image, it first divides the medical image into several pathological partitions, and then identifies each pathological partition to find abnormal points (lesions) and generates corresponding partition analysis information. In one example, the partition analysis information can be expressed as: there is an abnormal black dot (analysis result) in the left chest C1 area (pathological partition) of a chest CT picture (medical image). In the partition analysis information, it can include a detailed description of the lesion site, such as the position (absolute position and relative position), size (area), shape, etc. of the "abnormal black dot".

[0033] S30. Semantically parse the pathological analysis information through a pathological analysis model to generate disease condition reference information.

[0034] Understandably, a large number of analysis suggestions put forward by doctors for historical medical images can be collected, paired and stored with the pathological analysis information of the historical medical images to form paired data. Training these paired data can generate a corresponding pathological analysis model. Here, the pathological analysis model can be a neural network model based on semantic analysis.

[0035] After the pathological analysis information is semantically parsed by the pathological analysis model, corresponding disease condition reference information can be generated. The disease condition reference information can include medical advice information provided by doctors such as medication information, medication precautions information, and disease condition development trend prediction information.

[0036] S40. Obtain a reference image corresponding to the disease condition reference information.

[0037] Understandably, the specific disease type can be extracted from the disease condition reference information, and then the reference image of this type of disease can be obtained. Here, the reference image can be a typical example or sample belonging to a certain disease type. One disease condition reference information can correspond to one or more disease types. At least one reference image can be obtained for each disease type. Therefore, the reference image includes at least one reference image.

[0038] S50. Compare the differences between the medical image and the reference image to generate a comparison result.

[0039] Understandably, the medical image and the reference image can be directly compared, the difference degree between them can be calculated, and a corresponding comparison result can be generated. If the reference image includes multiple reference images, the medical image will be compared with each reference image respectively to generate a comparison result containing multiple image difference degrees. The comparison result can reflect the difference between the current medical image and the reference image, facilitating doctors to analyze whether the pathological analysis information and / or the disease condition reference information is reasonable.

[0040] S60. Import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0041] Understandably, the pathological analysis information, the disease condition reference information, and the comparison result can be filled into a preset reference information template to form the auxiliary reference information of the medical image. Different reference information templates can be set, such as the doctor version and the patient version, and then different versions of the auxiliary reference information can be generated. The specific reference information template can be set according to actual needs and will not be elaborated here. In some examples, the auxiliary reference information can also be represented in the form of a chart. The auxiliary reference information is based on model recognition (including two models) and the result of historical data comparison, which can help doctors or patients quickly understand the information contained in the medical image and reduce the misreading of the medical image.

[0042] In steps S10 - S60, the medical image to be parsed is obtained to identify the corresponding pathological analysis information through the medical image. The medical image is processed by an image parsing model to generate pathological analysis information including several pathological partitions. Here, by parsing the medical image in the way of pathological partitions, fine local pathological information can be obtained. The pathological analysis information is semantically parsed by a pathological analysis model to generate disease condition reference information. Here, by using semantic recognition to interpret the pathological analysis information, more accurate disease condition reference information can be obtained. The reference image corresponding to the disease condition reference information is obtained. Here, the reference image can reflect the typical characteristics of a certain disease type. The difference between the medical image and the reference image is compared to generate a comparison result. Here, the comparison result can reflect the approximation degree of the two images. The pathological analysis information, the disease condition reference information, and the comparison result are imported into the reference information template to generate the auxiliary reference information of the medical image. Here, the generated auxiliary reference information is richer and more accurate. In this embodiment, two neural network models are used to parse the medical image in stages, and at the same time, the reference image is used to verify the medical image, ensuring the accuracy of the pathological analysis information and the disease condition reference information. This embodiment improves the recognition ability of the medical image and can provide more accurate reference opinions for medical personnel.

[0043] Optionally, step S20, that is, processing the medical image by the image parsing model to generate pathological analysis information, the pathological analysis information includes several partition parsing information, the medical image is divided into several pathological partitions, and one pathological partition corresponds to one partition parsing information, includes:

[0044] S201. Divide the medical image into several pathological partitions according to a preset specification;

[0045] S202. Perform feature recognition from the pathological partitions to extract partition features;

[0046] S203. Input several of the partition features into the image parsing model to obtain the partition parsing information output by the image parsing model, where the partition parsing information includes the lesion location, lesion degree, and associated lesion information;

[0047] S204. Fill several of the partition parsing information into a preset analysis template to generate the pathological analysis information.

[0048] Understandably, the preset specifications can be set according to actual needs. Medical images can be divided into several pathological partitions according to the preset specifications. Different pathological partitions can partially overlap or be separated from each other. Several partition features can be extracted from the pathological partitions. Several partition features can be extracted from one pathological partition. In some cases, one partition feature can be represented by a multi-dimensional feature vector.

[0049] The image parsing model can be a neural network model trained based on a large number of historical medical images. By processing the partition features of the pathological partitions through the image parsing model, partition parsing information corresponding to the pathological partitions can be generated. The partition parsing information includes the characteristic data of the lesions in the pathological partitions, such as the lesion location, lesion degree, and associated lesion information, etc.

[0050] Fill several partition parsing information into a preset analysis template to generate the pathological analysis information. Different preset analysis templates can be set, and then different versions of the pathological analysis information can be generated. The specific preset analysis template can be set according to actual needs and will not be elaborated here. The pathological analysis information can be the sum of the partition parsing information of all pathological partitions. In some examples, the pathological analysis information also includes the associated lesion information between different pathological partitions.

[0051] Optionally, step S204, that is, filling several of the partition parsing information into a preset analysis template to generate the pathological analysis information, includes:

[0052] S2041. Determine whether the amount of information contained in the partition parsing information is greater than a preset information threshold;

[0053] S2042. If it is greater, divide the pathological partition corresponding to the partition parsing information into multiple sub-pathological partitions;

[0054] S2043. Extract sub-partition features from the sub-pathological partitions;

[0055] S2044. Process the sub-partition features of the sub-pathological partitions through a sub-model of the image parsing model to generate sub-partition parsing information corresponding to the sub-pathological partitions;

[0056] S2045. Generate the pathological analysis information according to the partition parsing information and the sub-partition parsing information.

[0057] Understandably, the amount of information contained in the partition analysis information can refer to the number or scale (area) of lesions. The preset information amount threshold can be set according to actual needs. For example, if the number or area of lesions in a certain pathological partition is greater than the corresponding information amount threshold, then the pathological partition needs to be sliced to form multiple sub-pathological partitions. If the number or area of lesions in a certain pathological partition is less than or equal to the corresponding information amount threshold, then there is no need to slice the pathological partition.

[0058] The sub-model of the image analysis model refers to the model specifically used to identify specific parts. For example, a model for the C1 area of the left chest can be constructed to identify the refined features of multiple sub-pathological partitions in the C1 area of the left chest. The pathological analysis information includes all partition analysis information and sub-partition analysis information.

[0059] Optionally, step S30, that is, processing the pathological analysis information through the pathological analysis model to generate the disease condition reference information, includes:

[0060] S301. Extract disease condition keyword groups from the pathological analysis information through a preset extraction mechanism, where the disease condition keyword groups include several disease condition keywords and several disease condition keyword combinations;

[0061] S302. Obtain a set of disease condition suggestions that match the disease condition keyword groups, where the set of disease condition suggestions includes multiple disease condition suggestions; each disease condition suggestion matches at least one of the disease condition keywords or the disease condition keyword combinations in the disease condition keyword groups;

[0062] S303. Screen the multiple disease condition suggestions in the set of disease condition suggestions through a preset screening mechanism to generate the disease condition reference information.

[0063] Understandably, the preset extraction mechanism can be set according to actual needs. The disease condition keyword groups include several disease condition keywords and several disease condition keyword combinations. The disease condition keywords can be single keywords, such as chest infection. The disease condition keyword combinations include at least two keywords, such as combination 1: chest infection, area XX.

[0064] After determining the disease condition keyword groups, corresponding disease condition suggestions can be matched through the disease condition keywords and disease condition keyword combinations in the disease condition keyword groups to form a set of disease condition suggestions. One disease condition keyword can match at least one disease condition suggestion.

[0065] The preset screening mechanism can be set according to actual needs. For example, for disease condition keyword 1, the correlation degrees of multiple disease condition suggestions matched by disease condition keyword 1 with other disease condition keywords can be calculated, and the disease condition suggestion with the highest correlation degree can be selected. Here, the pathological analysis model includes the above preset extraction mechanism and preset screening mechanism.

[0066] Optionally, step S40, that is, obtaining a reference image corresponding to the disease reference information, includes:

[0067] S401. Extract at least one key information from the disease reference information;

[0068] S402. Obtain a reference image corresponding to the key information, and the reference image includes at least one of the reference images.

[0069] Understandably, key information refers to information that can reflect a specific disease condition. In one example, a key information can be the type of a certain lung infection. The reference image corresponding to the key information can be an infection image of a typical case of the same disease type, or an image fitted by software.

[0070] One or more key information can be extracted from the disease reference information. Each key information can match at least one reference image. The reference image refers to all the reference images matched by all the key information.

[0071] Optionally, step S50, that is, comparing the differences between the medical image and the reference image to generate a comparison result;

[0072] S501. Obtain a plurality of characteristic indicators associated with the key information;

[0073] S502. Calculate the characteristic difference degree between the medical image and the reference image on the characteristic indicators;

[0074] S503. Determine the comparison result of the specified dimension between the medical image and the reference image according to a plurality of the characteristic difference degrees, and the comparison result includes at least one of the comparison results of the specified dimension.

[0075] Understandably, the characteristic indicators associated with different key information generally have differences. For example, the characteristic indicators associated with lung infection are different from those associated with gastrointestinal infection.

[0076] After determining a plurality of characteristic indicators, the characteristic difference degree between the medical image and the reference image on the characteristic indicators can be calculated. The specific calculation method of the characteristic difference degree can be set according to actual needs. For example, it can be an absolute value or a percentage; it can be based on the medical image or the reference image, etc.

[0077] The comparison result of the specified dimension refers to the comparison result determined based on the dimension of the current key information. The comparison result of the specified dimension can be the weighted sum of a plurality of characteristic difference degrees. One key information corresponds to one comparison result of the specified dimension. If there are multiple key information, the comparison result includes multiple comparison results of the specified dimension.

[0078] Optionally, step S60, namely, importing the pathological analysis information, the condition reference information and the comparison result into a reference information template to generate auxiliary reference information of the medical image, includes:

[0079] S601: If the comparison result shows that the characteristic difference between the medical image and the reference image in the characteristic index is greater than or equal to a preset difference threshold, a question mark is added to the pathological analysis information and / or the condition reference information corresponding to the key information;

[0080] S602. If the comparison result shows that the feature difference between the medical image and the reference image in the feature index is less than a preset difference threshold, a credible mark is added to the pathological analysis information and / or the condition reference information corresponding to the key information; the auxiliary reference information includes the questionable mark and / or the credible mark.

[0081] Understandably, the preset difference threshold can be set according to actual conditions, such as 80%, 90%, etc. When the characteristic difference between the medical image and the reference image in the characteristic index is greater than or equal to the preset difference threshold, it indicates that the current pathological analysis information and / or condition reference information may be different from the actual result of the medical image, so a doubt mark can be added to facilitate the doctor to compare the result.

[0082] When the feature difference between the medical image and the reference image in terms of feature indicators is less than the preset difference threshold, it indicates that the current pathological analysis information and / or condition reference information has a high degree of match with the actual results of the medical image and is not prone to errors. Therefore, a reliable mark can be added to reduce the doctor's comparison time.

[0083] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0084] In one embodiment, a device for generating medical auxiliary information based on image analysis is provided, and the device for generating medical auxiliary information based on image analysis corresponds to the method for generating medical auxiliary information based on image analysis in the above-mentioned embodiment. Figure 3 As shown, the medical auxiliary information generation device based on image analysis includes an acquisition module 10, an image analysis module 20, a semantic analysis module 30, a reference image acquisition module 40, an image comparison module 50 and a reference information generation module 60. The functional modules are described in detail as follows:

[0085] An acquisition module 10, used to acquire the medical image to be analyzed;

[0086] An image analysis module 20, configured to process the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes a plurality of partition analysis information, the medical image is divided into a plurality of pathological partitions, and one of the pathological partitions corresponds to one of the partition analysis information;

[0087] A semantic analysis module 30, configured to perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information;

[0088] A reference image acquisition module 40, configured to acquire a reference image corresponding to the disease condition reference information;

[0089] An image comparison module 50, configured to compare the differences between the medical image and the reference image to generate a comparison result;

[0090] A reference information generation module 60, configured to import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0091] Optionally, the image analysis module 20 includes:

[0092] A region division unit, configured to divide the medical image into a plurality of pathological partitions according to a preset specification;

[0093] A partition feature extraction unit, configured to perform feature recognition on the pathological partition to extract partition features;

[0094] A partition analysis information generation unit, configured to input a plurality of the partition features into the image analysis model to obtain the partition analysis information output by the image analysis model, where the partition analysis information includes lesion location, lesion degree, and associated lesion information;

[0095] A pathological analysis information generation unit, configured to fill a plurality of the partition analysis information into a preset analysis template to generate the pathological analysis information.

[0096] Optionally, the pathological analysis information generation unit includes:

[0097] An information amount judgment unit, configured to judge whether the information amount included in the partition analysis information is greater than a preset information amount threshold;

[0098] A region division unit, configured to, if it is greater, divide the pathological partition corresponding to the partition analysis information into a plurality of sub-pathological partitions;

[0099] A sub-partition feature extraction unit, configured to extract sub-partition features from the sub-pathological partitions;

[0100] A sub - partition parsing unit, configured to process the sub - partition features of the sub - pathological partition through a sub - model of the image parsing model, and generate sub - partition parsing information corresponding to the sub - pathological partition;

[0101] An analysis information generation unit, configured to generate the pathological analysis information according to the partition parsing information and the sub - partition parsing information.

[0102] Optionally, the semantic parsing module 30 includes:

[0103] An extraction unit, configured to extract a disease keyword group from the pathological analysis information through a preset extraction mechanism, where the disease keyword group includes a plurality of disease keywords and a plurality of disease keyword combinations;

[0104] A disease advice set acquisition unit, configured to acquire a disease advice set matching the disease keyword group, where the disease advice set includes multiple disease advices; each of the disease advices matches at least one of the disease keywords or the disease keyword combinations in the disease keyword group;

[0105] A screening unit, configured to screen the multiple disease advices in the disease advice set through a preset screening mechanism to generate the disease reference information.

[0106] Optionally, the reference image acquisition module 40 includes:

[0107] A key information extraction unit, configured to extract at least one piece of key information from the disease reference information;

[0108] A reference image matching unit, configured to acquire a reference image corresponding to the key information, where the reference image includes at least one of the reference images.

[0109] Optionally, the image comparison module 50 includes;

[0110] A feature index acquisition unit, configured to acquire a plurality of feature indexes associated with the key information;

[0111] A difference degree calculation unit, configured to calculate the feature difference degree between the medical image and the reference image in the feature indexes;

[0112] A comparison result determination unit, configured to determine the comparison result of the specified dimension between the medical image and the reference image according to the plurality of feature difference degrees, where the comparison result includes at least one of the specified dimension comparison results.

[0113] Optionally, the reference information generation module 60 includes:

[0114] Add a doubtful identification unit for adding a doubtful identification to the pathological analysis information and / or the disease condition reference information corresponding to the key information if the comparison result indicates that the feature difference degree between the medical image and the reference image in the feature index is greater than or equal to a preset difference degree threshold;

[0115] Add a credible identification unit for adding a credible identification to the pathological analysis information and / or the disease condition reference information corresponding to the key information if the comparison result indicates that the feature difference degree between the medical image and the reference image in the feature index is less than a preset difference degree threshold; the auxiliary reference information includes the doubtful identification and / or the credible identification.

[0116] For the specific limitations of the medical auxiliary information generation device based on image analysis, reference can be made to the limitations of the medical auxiliary information generation method based on image analysis in the above text, which will not be elaborated here. Each module in the above medical auxiliary information generation device based on image analysis can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0117] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 4 shown. The computer device includes a processor, a memory, a network interface, and a database 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 readable storage medium and an internal memory. The readable storage medium stores an operating system, computer-readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the readable storage medium. The database of the computer device is used to store the data involved in the medical auxiliary information generation method based on image analysis. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, a medical auxiliary information generation method based on image analysis is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0118] In one embodiment, a computer device is provided, including a memory, a processor, and computer-readable instructions stored on the memory and executable on the processor. When the processor executes the computer-readable instructions, the following steps are implemented:

[0119] Obtain a medical image to be analyzed;

[0120] Process the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one pathological partition corresponds to one partition analysis information;

[0121] Perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information;

[0122] Obtain a reference image corresponding to the disease condition reference information;

[0123] Compare the differences between the medical image and the reference image to generate a comparison result;

[0124] Import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0125] In one embodiment, one or more computer-readable storage media storing computer-readable instructions are provided. The readable storage media provided in this embodiment include non-volatile readable storage media and volatile readable storage media. Computer-readable instructions are stored on the readable storage media, and when the computer-readable instructions are executed by one or more processors, the following steps are implemented:

[0126] Obtain a medical image to be analyzed;

[0127] Process the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one pathological partition corresponds to one partition analysis information;

[0128] Perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information;

[0129] Obtain a reference image corresponding to the disease condition reference information;

[0130] Compare the differences between the medical image and the reference image to generate a comparison result;

[0131] Import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate auxiliary reference information for the medical image.

[0132] 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 computer-readable instructions. The computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above 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 non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0133] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0134] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for generating medical auxiliary information based on image analysis, characterized in that, it includes: Obtain the medical image to be analyzed; Process the medical image through an image analysis model to generate pathological analysis information, where the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one of the pathological partitions corresponds to one of the partition analysis information; Perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information; Obtain a reference image corresponding to the disease condition reference information; Compare the differences between the medical image and the reference image to generate a comparison result; Import the pathological analysis information, the disease condition reference information, and the comparison result into a reference information template to generate the auxiliary reference information of the medical image; The pathological analysis model includes a preset extraction mechanism and a preset screening mechanism; The process of generating disease condition reference information by processing the pathological analysis information through the pathological analysis model includes: Extract disease condition keyword groups from the pathological analysis information through a preset extraction mechanism, where the disease condition keyword groups include several disease condition keywords and several disease condition keyword combinations; Obtain a set of disease condition suggestions that match the disease condition keyword groups, where the set of disease condition suggestions includes multiple disease condition suggestions; each of the disease condition suggestions matches at least one of the disease condition keywords or the disease condition keyword combinations in the disease condition keyword groups; Screen the multiple disease condition suggestions in the set of disease condition suggestions through a preset screening mechanism to generate the disease condition reference information.

2. The method for generating medical auxiliary information based on image analysis according to claim 1, characterized in that, The process of generating pathological analysis information by processing the medical image through an image analysis model, where the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one of the pathological partitions corresponds to one of the partition analysis information, includes: Divide the medical image into several pathological partitions according to a preset specification; Perform feature recognition on the pathological partitions to extract partition features; Input several of the partition features into the image analysis model to obtain the partition analysis information output by the image analysis model, where the partition analysis information includes lesion location, lesion degree, and associated lesion information; Fill several of the partition analysis information into a preset analysis template to generate the pathological analysis information.

3. The method for generating medical auxiliary information based on image analysis according to claim 2, characterized in that, The process of filling several of the partition analysis information into a preset analysis template to generate the pathological analysis information includes: Judge whether the amount of information contained in the partition analysis information is greater than a preset information amount threshold; If it is greater, divide the pathological partition corresponding to the partition analysis information into multiple sub-pathological partitions; Extract sub-partition features from the sub-pathological partitions; Process the sub-partition features of the sub-pathological partitions through a sub-model of the image analysis model to generate sub-partition analysis information corresponding to the sub-pathological partitions; Generate the pathological analysis information according to the partition analysis information and the sub-partition analysis information.

4. The method for generating medical assistance information based on image analysis as claimed in claim 1, wherein, the obtaining of the reference image corresponding to the disease condition reference information includes: extracting at least one key information from the disease condition reference information; obtaining a reference image corresponding to the key information, and the reference image includes at least one of the reference images.

5. The method for generating medical assistance information based on image analysis as claimed in claim 4, wherein, the comparing of the differences between the medical image and the reference image to generate a comparison result includes: obtaining a plurality of characteristic indexes associated with the key information; calculating the characteristic difference degree between the medical image and the reference image on the characteristic indexes; determining the comparison result of the specified dimension between the medical image and the reference image according to a plurality of the characteristic difference degrees, and the comparison result includes at least one of the comparison results of the specified dimension.

6. The method for generating medical assistance information based on image analysis as claimed in claim 5, wherein, the importing of the pathological analysis information, the disease condition reference information and the comparison result into a reference information template to generate the auxiliary reference information of the medical image includes: if the comparison result indicates that the characteristic difference degree between the medical image and the reference image on the characteristic indexes is greater than or equal to a preset difference degree threshold, adding a doubtful mark to the pathological analysis information and / or the disease condition reference information corresponding to the key information; if the comparison result indicates that the characteristic difference degree between the medical image and the reference image on the characteristic indexes is less than the preset difference degree threshold, adding a credible mark to the pathological analysis information and / or the disease condition reference information corresponding to the key information; the auxiliary reference information includes the doubtful mark and / or the credible mark.

7. A device for generating medical assistance information based on image analysis, wherein, it includes: an obtaining module, configured to obtain a medical image to be analyzed; an image analysis module, configured to process the medical image through an image analysis model to generate pathological analysis information, the pathological analysis information includes several partition analysis information, the medical image is divided into several pathological partitions, and one of the pathological partitions corresponds to one of the partition analysis information; a semantic analysis module, configured to perform semantic analysis on the pathological analysis information through a pathological analysis model to generate disease condition reference information; a reference image obtaining module, configured to obtain a reference image corresponding to the disease condition reference information; an image comparison module, configured to compare the differences between the medical image and the reference image to generate a comparison result; a reference information importing module, configured to import the pathological analysis information, the disease condition reference information and the comparison result into a reference information template to generate the auxiliary reference information of the medical image; the semantic analysis module includes: an extraction unit, configured to extract disease condition keyword groups from the pathological analysis information through a preset extraction mechanism, and the disease condition keyword groups include several disease condition keywords and several disease condition keyword combinations; An advice collection obtaining unit, configured to obtain a disease advice collection that matches the disease keyword group, where the disease advice collection includes multiple disease advices; each of the disease advices matches at least one of the disease keywords or the disease keyword combinations in the disease keyword group; A screening unit, configured to screen multiple disease advices in the disease advice collection through a preset screening mechanism to generate the disease reference information.

8. A computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, characterized in that, when the processor executes the computer-readable instructions, the method for generating medical auxiliary information based on image parsing according to any one of claims 1 to 6 is implemented.

9. One or more readable storage media storing computer-readable instructions, when the computer-readable instructions are executed by one or more processors, enabling the one or more processors to execute the method for generating medical auxiliary information based on image parsing according to any one of claims 1 to 6.

Images