Medical report generation method and device, electronic equipment and storage medium

Abstract

The invention provides a medical report generation method and device, electronic equipment and a storage medium. Comprising the following steps: acquiring a medical image with annotation information, and preprocessing the medical image; carrying out abnormal region identification on the preprocessed medical image to obtain a segmented image of an abnormal region; generating a standard pathological text based on the segmentation map of the abnormal region, the annotation information and a clinical record knowledge base; and generating a medical report based on the standard pathological text, the segmentation map of the abnormal region and the medical report template. By adopting the technical scheme, the problems of low generation efficiency of the medical report and non-uniform quality of the medical report in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of computer medicine technology, and in particular to a method, device, electronic equipment and storage medium for generating a medical report. Background Art

[0002] With the rapid development of medical imaging technology, imaging methods such as CT and ultrasound are increasingly used in clinical diagnosis. Doctors can observe and analyze patients' lesions through these images and issue medical reports. At present, the generation of medical reports mainly relies on manual operations of doctors, who need to manually edit medical reports based on patients' medical images. The generation efficiency of medical reports is low, and the diagnostic results in medical reports may be biased due to human factors. Summary of the invention

[0003] The present invention provides a method, device, electronic device and storage medium for generating a medical report, which are used to solve the problems of low generation efficiency and inconsistent quality of medical reports in the prior art.

[0004] In a first aspect, the present invention provides a method for generating a medical report, comprising: obtaining a medical image with annotation information, and preprocessing the medical image. Performing abnormal region identification on the preprocessed medical image to obtain a segmentation map of the abnormal region. Generating a standard pathology text based on the segmentation map of the abnormal region, the annotation information, and a clinical record knowledge base. Generating a medical report based on the standard pathology text, the segmentation map of the abnormal region, and a medical report template.

[0005] Optionally, a standard pathology text is generated based on the segmentation map of the abnormal region, the annotation information, and the clinical record knowledge base, including determining the characteristics of the lesion in the segmentation map of the abnormal region based on the segmentation map of the abnormal region and the annotation information, the characteristics of the lesion including the type of the lesion, the size of the lesion, and the location of the lesion. An initial pathology text is generated based on the characteristics of the lesion, the annotation information, and the clinical record knowledge base. A standard pathology text is generated based on a predefined text template and the initial pathology text.

[0006] Optionally, determining the characteristics of the lesion in the segmentation map of the abnormal region based on the segmentation map and the annotation information of the abnormal region includes: using the segmentation map and the annotation information of the abnormal region to query a pathology knowledge base to obtain a pathology map that matches the segmentation map and the annotation information of the abnormal region. According to the description information of the pathology map, determining the characteristics of the lesion in the segmentation map of the abnormal region.

[0007] Optionally, after generating a medical report based on the standard pathology text, the segmentation map of the abnormal area, and the medical report template, the method further includes obtaining a doctor's review result of the medical report, and judging whether the medical report is an accurate medical report based on the review result. If the medical report is an accurate medical report, the accurate medical report is saved as the patient's official medical report. If the medical report is an inaccurate medical report, identification information is added to the inaccurate medical report.

[0008] Optionally, generating a medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template includes obtaining patient information corresponding to the medical image and generating the medical report template based on the patient information, wherein the patient information includes the patient's name, the patient's age, the examination number, the examination date and the examination department. The standard pathology text and the segmentation map of the abnormal area are added to the corresponding positions of the medical report template to obtain the medical report.

[0009] Optionally, the annotation information is organ position information corresponding to the medical image.

[0010] In a second aspect, the present invention provides a device for generating a medical report, including: a first processing module for obtaining a medical image with annotation information and preprocessing the medical image; an identification module for identifying abnormal regions of the preprocessed medical image to obtain a segmentation map of the abnormal region; a first generation module for generating a standard pathology text based on the segmentation map of the abnormal region, the annotation information, and a clinical record knowledge base; and a second generation module for generating a medical report based on the standard pathology text, the segmentation map of the abnormal region, and a medical report template.

[0011] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for generating a medical report as described above is implemented.

[0012] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for generating a medical report.

[0013] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which implements any of the above-mentioned methods for generating a medical report when executed by a processor.

[0014] The method, device, electronic device and storage medium for generating medical reports provided by the present invention obtain medical images with annotation information and preprocess these images to ensure the quality and consistency of the medical images. Then, the abnormal area of ​​the preprocessed medical image is identified to obtain a segmentation map of the abnormal area. In this way, accurate positioning of the disease characteristic area is achieved. Based on the segmentation map of the abnormal area, the annotation information and the clinical record knowledge base, a pathological text that meets the standards is generated to ensure the integrity and accuracy of the pathological text. A medical report is generated by combining the standard pathological text, the segmentation map of the abnormal area and the preset medical report template. The efficiency of generating medical reports is improved through an automated report generation method, and the quality of medical reports is improved by combining the clinical record knowledge base, thereby avoiding quality problems in medical reports caused by human factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flowchart of the method for generating a medical report provided by the present invention.

[0017] Figure 2 It is a flow chart of the respiratory report generation system provided by the present invention.

[0018] Figure 3 It is a schematic diagram of the structure of the device for generating a medical report provided by the present invention.

[0019] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Figure 1 FIG. 1 is a flow chart of the method for generating a medical report provided by the present invention. Figure 1 As shown, the method includes the following steps S101 to S104.

[0022] Step S101: Obtain a medical image with annotation information, and pre-process the medical image.

[0023] In an exemplary embodiment, the annotation information is the organ position information corresponding to the medical image.

[0024] It should be noted that medical images are images of the internal structure of the human body obtained through various medical imaging technologies (such as X-ray, CT, MRI, ultrasound, etc.).

[0025] As an optional embodiment, the medical image may be an endoscopic image.

[0026] As an optional embodiment, assuming that the medical image is a chest CT image, the annotation information may be the position of the left and right lungs, the position of the heart, the distribution of the ribs in the chest cavity, and the direction and branches of the trachea. The annotation information may help the medical report generation device or the multimodal large model to locate and perform pathological analysis more quickly.

[0027] As an optional embodiment, the annotation information may be added by the doctor after acquiring the medical image.

[0028] As an optional embodiment, the preprocessing includes denoising, standardizing and enhancing the medical image to improve the imaging quality of the medical image for subsequent analysis and processing. The present invention does not further limit the preprocessing method.

[0029] Step S102: performing abnormal region identification on the preprocessed medical image to obtain a segmentation map of the abnormal region.

[0030] As an optional embodiment, a deep learning segmentation model can be used to identify and segment abnormal areas of preprocessed medical images, such as airway abnormalities, mucosal lesions, neoplasms, etc.

[0031] Step S103: Generate a standard pathology text based on the segmentation map of the abnormal area, the annotation information and the clinical record knowledge base.

[0032] As an optional embodiment, the annotation information can assist the medical report generation device or the multimodal large model to more quickly locate the organ position or region position of the segmentation map of the abnormal region. For example, if the annotation information is trachea, the medical report generation device or the multimodal large model can directly search for pathological information related to trachea in the clinical record knowledge base, thereby improving the accuracy and efficiency of generating standard pathological texts.

[0033] As an optional embodiment, the clinical record knowledge base mainly contains the possible lesion types and clinical significance of abnormal areas. By calling verified pathology and diagnostic information from the clinical knowledge base, the generated pathology text can be based on a large number of historical cases, medical literature and standard diagnostic guidelines, improving the reliability of standard pathology texts. Through automated text generation, omissions or other possible errors in the manual input process are reduced, ensuring the integrity and standardization of the report.

[0034] In an exemplary embodiment, the above step S103 can be implemented by the following steps S11 to S13.

[0035] Step S11 : determining the features of the lesion in the segmentation map of the abnormal region based on the segmentation map and the annotation information of the abnormal region, wherein the features of the lesion include the type, size and location of the lesion.

[0036] As an optional embodiment, the medical report generation device or the multimodal large model preliminarily determines the characteristics of the lesion in the segmentation map of the abnormal area in combination with the annotation information. Specifically, the type of lesion in the segmentation map of the abnormal area is first identified. Then, the size of the lesion is measured through image processing technology, and the accurate anatomical location of the lesion is determined based on the annotation information, such as the left upper lobe of the lung, the right liver, etc., to ensure that the characteristics of the lesion are fully described.

[0037] Step S12: generating an initial pathology text based on the characteristics of the lesion, the annotation information and the clinical record knowledge base.

[0038] As an optional embodiment, the clinical record knowledge base contains a large number of verified pathology description templates and medical terms, calls relevant case records based on the characteristics of the lesion, and generates an initial pathology text that matches the current lesion condition.

[0039] Step S13: Generate a standard pathology text based on the predefined text template and the initial pathology text.

[0040] As an optional embodiment, the predefined text template includes a structured format and standardized terms of a standard pathology text. The pathological description of the lesion and the possible lesion type and clinical significance of the abnormal area are generated into a standard pathology text according to the predefined text template. The existing standard pathology text first describes the pathology and then states the possible lesion type and clinical significance.

[0041] In the above embodiment, the accuracy and comprehensiveness of the lesion description are ensured by extracting the lesion features, and a more comprehensive and accurate initial pathology text is generated by combining the clinical record knowledge base. The reliability of the pathology text is significantly improved by the automated standard pathology text, avoiding the problem of inconsistent pathology text quality due to differences in doctor levels.

[0042] In an exemplary embodiment, the above step S11 can be implemented by the following steps S111 to S112.

[0043] Step S111: Use the segmentation map and annotation information of the abnormal region to query the pathology knowledge base to obtain a pathology map that matches the segmentation map and annotation information of the abnormal region.

[0044] As an optional embodiment, the medical report generation device or multimodal large model uses the segmentation map and annotation information of the abnormal area to automatically query the pathology map matching the segmentation map and annotation information from the pathology knowledge base. The matching algorithm can be used to compare the abnormal area features (such as shape, density, and position) in the image with the existing pathology maps in the pathology knowledge base to select the most similar pathology maps. These pathology maps come from known cases and contain detailed pathology description information.

[0045] Step S112: Determine the characteristics of the lesion in the segmentation map of the abnormal region according to the description information of the pathological map.

[0046] As an optional embodiment, the medical report generation device or the multimodal large model determines the lesion features of the segmentation map of the abnormal region according to the description information in the matched pathology map. These lesion features are obtained from the description information in the pathology map.

[0047] In the above embodiment, the efficiency and accuracy of lesion feature recognition are significantly improved by automatically querying the pathology knowledge base. A large amount of case data in the pathology knowledge base is used to provide a reference for the segmentation map of each abnormal area, ensuring that the medical report generation device can make a scientific judgment based on the known clinical pathology results. The characteristics of the lesion in the segmentation map of the current abnormal area are determined based on the description information of the matched pathology map, which reduces the error of human subjective analysis and improves the consistency of diagnosis.

[0048] Step S104: Generate a medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template.

[0049] In an exemplary embodiment, the above step S104 can be implemented by the following steps S21-S22.

[0050] Step S21, obtaining patient information corresponding to the medical image, and generating a medical report template based on the patient information, the patient information including the patient's name, patient age, examination number, examination date and examination department.

[0051] As an optional embodiment, the medical report template predefines structured fields for patient information and image analysis content, and leaves space for inserting standard pathology text and segmentation map of abnormal areas.

[0052] Step S22: Add the standard pathology text and the segmentation map of the abnormal area to the corresponding positions of the medical report template to obtain a medical report.

[0053] As an optional embodiment, the generated standard pathology text and the segmentation map of the abnormal area are automatically added to the corresponding positions of the medical report template. The standard pathology text includes a detailed description of the lesion type, size and location, while the segmentation map of the abnormal area provides a visual lesion image.

[0054] In the above embodiment, by automatically generating a medical report template, the workload of the doctor manually filling in the basic information of the patient is reduced, and by automatically embedding the standard pathology text and the segmentation map of the abnormal area into the medical report template, the efficiency of generating the medical report is improved.

[0055] Through the above steps S101-S104, by obtaining medical images with annotation information and preprocessing these images, the quality and consistency of the medical images are ensured. Then, the abnormal area of ​​the preprocessed medical image is identified to obtain a segmentation map of the abnormal area. In this way, the precise positioning of the disease characteristic area is achieved. Based on the segmentation map of the abnormal area, the annotation information and the clinical record knowledge base, a pathological text that meets the standards is generated to ensure the integrity and accuracy of the pathological text. A medical report is generated by combining the standard pathological text, the segmentation map of the abnormal area and the preset medical report template. The efficiency of medical report generation is improved through the automated report generation method, and the quality of the medical report is improved by combining the clinical record knowledge base, avoiding quality problems in the medical report due to human factors.

[0056] In an exemplary embodiment, after generating a medical report based on the standard pathology text, the segmentation map of the abnormal region and the medical report template, the method further includes the following steps S31 to S33.

[0057] Step S31: Obtain the doctor's review result of the medical report, and determine whether the medical report is an accurate medical report based on the review result.

[0058] As an optional embodiment, the doctor checks whether there are errors or omissions by reviewing the patient information, standard pathology text and segmented map of abnormal areas in the report. Based on the doctor's feedback, it is automatically determined whether the medical report is an accurate medical report.

[0059] Step S32: If the medical report is an accurate medical report, save the accurate medical report as the patient's official medical report.

[0060] As an optional embodiment, if the medical report is confirmed to be correct, the accurate medical report is saved in the hospital information system as the patient's official medical report.

[0061] Step S33: If the medical report is an inaccurate medical report, add identification information to the inaccurate medical report.

[0062] As an optional embodiment, if the medical report is determined to be inaccurate, identification information is added to the report to prompt the doctor to pay special attention and decide whether it needs to be revised or regenerated. The identification information can be marked information such as "needs to be reviewed" or "to be revised".

[0063] In the above embodiment, the accuracy of the medical report is ensured by reviewing and re-reviewing by doctors, and adding identification information to inaccurate reports helps to mark medical reports that need to be revised and avoid misuse of erroneous reports.

[0064] The executor of each step in the above method can be a medical report generation device or a multimodal large model. The device can be implemented by software and / or hardware. The device can be integrated in an electronic device. The electronic device can be a terminal device (such as a smart phone, a personal computer, etc.), or a server (such as a local server or a cloud server, or a server cluster, etc.), or a processor, or a chip, etc.

[0065] Obviously, the above-described embodiments are only part of the embodiments of the present invention, not all of the embodiments. In order to better understand the above method, the above process is described below in conjunction with the embodiments, but the technical solutions of the embodiments of the present invention are not limited. Specifically,

[0066] The present invention provides a respiratory system endoscopic image analysis and report generation method based on artificial intelligence. Figure 2 1 is a flow chart of the respiratory report generation system provided by the present invention. The flow chart includes the following steps.

[0067] 1. Obtain the endoscopic image annotated by the doctor (i.e., the medical image in the above embodiment) through the API interface of the medical information management system. The format of the endoscopic image data includes but is not limited to JPEG and PNG.

[0068] 2. Use a large multimodal model to analyze the input endoscopic images. It should be noted that before the endoscopic images are input into the large multimodal model, the images need to be preprocessed, including denoising, standardization, and enhancement. In addition, a deep learning segmentation model is also needed to identify and segment abnormal areas of the preprocessed images, such as airway abnormalities, mucosal lesions, neoplasms, etc., to generate segmentation maps of abnormal areas.

[0069] It should be noted that this large multimodal model is trained by tens of thousands of real clinical respiratory report images and texts, and has the ability to process image and text data at the same time. It is a neural network model. The input of this large multimodal model is the segmentation map of abnormal areas in the endoscopic image, and the output is a medical report.

[0070] 3. Automatically generate structured medical reports based on image analysis results. Specific.

[0071] First, based on the segmentation map of the abnormal area, the prompt generation module of the large multimodal model combines the respiratory system professional knowledge base (i.e., the pathological knowledge base in the above embodiment, where the respiratory system professional knowledge base is a branch of the pathological knowledge base) and the historical clinical records (clinical record knowledge base) to generate preliminary text prompts (i.e., the initial pathological text in the above embodiment). Using the predefined text template, a standard pathological text is generated. It should be noted that the standard pathological text includes the specific location, morphological characteristics, possible lesion types and clinical significance of the abnormal area.

[0072] Then, a respiratory system endoscopy examination report (ie, the medical report in the above embodiment) that meets medical standards is automatically generated by the multimodal model based on standard pathological text and medical report template.

[0073] 4. After the report is generated, the results are fed back to the doctor through the user interface (feedback forms include text reports, annotated images, highlighting of key findings, etc.), and the doctor checks the generated report and image analysis results and compares them with the original images to verify whether the abnormalities identified by the large multimodal model are accurate.

[0074] 4-1. If the generated respiratory report meets the clinical standards, confirm and save it as the patient's official medical report.

[0075] 4-2. If the generated respiratory report does not meet the clinical standards, the doctor will modify or supplement it until it meets the requirements and save it as a formal report.

[0076] 5. Further training and optimization of large multimodal models based on generated medical reports and doctor feedback.

[0077] By automatically analyzing endoscopic images and quickly generating preliminary reports, diagnostic efficiency has been greatly improved, allowing doctors to focus more on the analysis of complex cases and patient communication. Using large multimodal models and deep learning technology, it can capture subtle abnormalities that the human eye may overlook. At the same time, through training with a large amount of clinical data, it provides more consistent and objective analysis results, thereby enhancing diagnostic accuracy. Secondly, the generation of standardized reports helps to standardize medical records and reduce the problem of inconsistent report quality caused by differences in personal habits.

[0078] The medical report generation device provided by the present invention is described below. The medical report generation device described below and the medical report generation method described above can refer to each other.

[0079] Figure 3 FIG. 1 is a schematic diagram of the structure of the device for generating a medical report provided by the present invention. Figure 3 As shown, the device includes:

[0080] The first processing module 310 is used to obtain the medical image with annotation information and pre-process the medical image.

[0081] The identification module 320 is used to identify abnormal regions in the preprocessed medical image to obtain a segmentation map of the abnormal regions.

[0082] The first generating module 330 is used to generate a standard pathology text based on the segmentation map of the abnormal region, the annotation information and the clinical record knowledge base.

[0083] The second generating module 340 is used to generate a medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template.

[0084] The above-mentioned device obtains medical images with annotation information and preprocesses these images to ensure the quality and consistency of the medical images. Then, the abnormal area of ​​the preprocessed medical image is identified to obtain a segmentation map of the abnormal area. In this way, the precise positioning of the disease characteristic area is achieved. Based on the segmentation map of the abnormal area, the annotation information and the clinical record knowledge base, a pathological text that meets the standards is generated to ensure the integrity and accuracy of the pathological text. A medical report is generated by combining the standard pathological text, the segmentation map of the abnormal area and the preset medical report template. The efficiency of medical report generation is improved through the automated report generation method, and the quality of the medical report is improved by combining the clinical record knowledge base, avoiding quality problems in the medical report caused by human factors.

[0085] Optionally, the first generation module 330 is further used to determine the characteristics of the lesion in the segmentation map of the abnormal region based on the segmentation map and the annotation information of the abnormal region, the characteristics of the lesion including the type of the lesion, the size of the lesion and the location of the lesion. An initial pathology text is generated based on the characteristics of the lesion, the annotation information and the clinical record knowledge base. A standard pathology text is generated based on a predefined text template and the initial pathology text.

[0086] Optionally, the first generation module 330 is further configured to query the pathology knowledge base using the segmentation map and annotation information of the abnormal region to obtain a pathology map that matches the segmentation map and annotation information of the abnormal region, and to determine the characteristics of the lesion in the segmentation map of the abnormal region according to the description information of the pathology map.

[0087] Optionally, the device further includes a second processing module, which is used to obtain the doctor's review result of the medical report after generating the medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template, and determine whether the medical report is an accurate medical report based on the review result. If the medical report is an accurate medical report, the accurate medical report is saved as the patient's official medical report. If the medical report is an inaccurate medical report, identification information is added to the inaccurate medical report.

[0088] Optionally, the second generation module 340 is further used to obtain patient information corresponding to the medical image, and generate a medical report template based on the patient information, wherein the patient information includes the patient's name, patient age, examination number, examination date, and examination department. The standard pathology text and the segmentation map of the abnormal area are added to the corresponding position of the medical report template to obtain a medical report.

[0089] Optionally, the annotation information is organ position information corresponding to the medical image.

[0090] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute a method for generating a medical report, which includes:

[0091] Obtain medical images with annotation information and preprocess the medical images.

[0092] The abnormal region is identified on the preprocessed medical image to obtain a segmentation map of the abnormal region.

[0093] Generate standard pathology text based on the segmentation map, annotation information and clinical record knowledge base of the abnormal area.

[0094] Generate medical reports based on standard pathology text, segmentation maps of abnormal areas, and medical report templates.

[0095] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk.

[0096] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the medical report generation method provided by the above-mentioned methods, which includes.

[0097] Obtain medical images with annotation information and preprocess the medical images.

[0098] The abnormal region is identified on the preprocessed medical image to obtain a segmentation map of the abnormal region.

[0099] Generate standard pathology text based on the segmentation map, annotation information and clinical record knowledge base of the abnormal area.

[0100] Generate medical reports based on standard pathology text, segmentation maps of abnormal areas, and medical report templates.

[0101] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method for generating a medical report provided by the above-mentioned methods, the method comprising.

[0102] Obtain medical images with annotation information and preprocess the medical images.

[0103] The abnormal region is identified on the preprocessed medical image to obtain a segmentation map of the abnormal region.

[0104] Generate standard pathology text based on the segmentation map, annotation information and clinical record knowledge base of the abnormal area.

[0105] Generate medical reports based on standard pathology text, segmentation maps of abnormal areas, and medical report templates.

[0106] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0107] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0108] Finally, it should be noted that 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 above embodiments, it should be understood by those skilled in the art that they can still modify the technical solutions described in the above embodiments, or replace some of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating a medical report, characterized in that: include: Acquire a medical image with annotation information, and preprocess the medical image; Perform abnormal region recognition on the preprocessed medical image to obtain a segmentation map of the abnormal region; Generate a standard pathology text based on the segmentation map of the abnormal area, the annotation information and a clinical record knowledge base; A medical report is generated based on the standard pathology text, the segmentation map of the abnormal area and a medical report template.

2. The method for generating a medical report according to claim 1, characterized in that: The generating of a standard pathology text based on the segmentation map of the abnormal region, the annotation information and the clinical record knowledge base includes: Determine the features of the lesion in the segmentation map of the abnormal region based on the segmentation map of the abnormal region and the annotation information, wherein the features of the lesion include the type of the lesion, the size of the lesion and the location of the lesion; generating an initial pathology text based on the characteristics of the lesion, the annotation information and the clinical record knowledge base; A standard pathology text is generated based on a predefined text template and the initial pathology text.

3. The method for generating a medical report according to claim 2, characterized in that: The determining the features of the lesion in the segmentation map of the abnormal region based on the segmentation map of the abnormal region and the annotation information comprises: Using the segmentation map of the abnormal region and the annotation information to query a pathology knowledge base, to obtain a pathology map that matches the segmentation map of the abnormal region and the annotation information; According to the description information of the pathological image, the characteristics of the lesion in the segmentation image of the abnormal area are determined.

4. The method for generating a medical report according to any one of claims 1 to 3, characterized in that: After generating a medical report based on the standard pathology text, the segmentation map of the abnormal region and the medical report template, the method further includes: Obtaining the doctor's review result of the medical report, and judging whether the medical report is an accurate medical report based on the review result; If the medical report is an accurate medical report, saving the accurate medical report as the official medical report of the patient; In the case that the medical report is an inaccurate medical report, identification information is added to the inaccurate medical report.

5. The method for generating a medical report according to any one of claims 1 to 3, characterized in that: The generating of a medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template comprises: Acquire patient information corresponding to the medical image, and generate a medical report template based on the patient information, wherein the patient information includes the patient's name, patient's age, examination number, examination date, and examination department; The standard pathology text and the segmentation map of the abnormal area are added to corresponding positions of the medical report template to obtain the medical report.

6. The method for generating a medical report according to any one of claims 1 to 3, characterized in that: The annotation information is the organ position information corresponding to the medical image.

7. A device for generating a medical report, characterized in that: include: A first processing module, used for acquiring a medical image with annotation information and preprocessing the medical image; The recognition module is used to identify abnormal regions of the preprocessed medical image and obtain a segmentation map of the abnormal region; A first generating module, configured to generate a standard pathology text based on the segmentation map of the abnormal region, the annotation information and a clinical record knowledge base; The second generating module is used to generate a medical report based on the standard pathology text, the segmentation map of the abnormal area and the medical report template.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for generating a medical report according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating a medical report according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for generating a medical report according to any one of claims 1 to 6 is implemented.

Images