Radiology report acquisition method and device, electronic equipment and storage medium
Through the combination of deep neural networks and medical knowledge graphs, the process of obtaining radiological reports is automated, which solves the problem of inefficient acquisition of radiological reports in the existing technology, and achieves efficient and reliable report generation.
Patent Information
- Application Number
- CN202411427685.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-14
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The process of obtaining radiological reports of existing radio equipment is cumbersome, consuming a lot of manpower and time resources, resulting in inefficiency.
Deep neural networks are used to combine medical knowledge graphs to automatically obtain radiographic descriptions, entity names and processing solutions, train models through cross-entropy loss functions, generate radiological reports, and reduce manual intervention.
It improves the efficiency of obtaining radiological reports, saves manpower and time resources, and improves the reliability of reports, reducing the impact of human factors.
Smart Images

Figure CN118969171A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of Internet technology, and in particular to a method, device, electronic device and storage medium for obtaining a radiology report. Background Art
[0002] With the continuous advancement of technology, radiation equipment has also shown broad application prospects in scientific research, teaching and health management. Radiation equipment includes X-ray machines, computer tomography machines, and magnetic resonance imaging machines. They have greatly improved the quality and efficiency of medical services and are an important symbol of the progress of modern medicine.
[0003] However, the process of obtaining the current radiology report of the existing radiological equipment is cumbersome, which is not conducive to improving the efficiency of obtaining the current radiology report. The reason is that the existing technology mainly adopts the manual writing method, and the manual writing takes a long time, so it will consume a lot of human resources and time resources. Therefore, the process of obtaining the current radiology report of the existing technology is cumbersome, which is not conducive to improving the efficiency of obtaining the current radiology report. Summary of the invention
[0004] The embodiments of the present application provide a method, device, electronic device and storage medium for obtaining a radiology report to solve the technical problem that the process of obtaining the current radiology report of the above-mentioned existing radiological equipment is cumbersome and is not conducive to improving the efficiency of obtaining the current radiology report.
[0005] In a first aspect, an embodiment of the present application provides a method for obtaining a radiology report, which is applied to an electronic device, wherein the electronic device stores a deep neural network, and the method for obtaining a radiology report includes: Acquire a preset radiographic image of a radiological device, and acquire a preset image description of the preset radiographic image; In a preset medical knowledge graph, obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description; Combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into a training sample; Training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; By using a preset cross entropy loss function, a first cross entropy loss value between the predicted image description and the preset image description is obtained, a second cross entropy loss value between the predicted entity name and the preset entity name is obtained, a third cross entropy loss value between the predicted entity description and the preset entity description is obtained, and a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme is obtained; Generate a current loss value corresponding to the training sample according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value generation model; If the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained.
[0006] Among them, the current loss value generation model is a generation model of the current loss value.
[0007] Exemplarily, the current loss value generation model is: T=W1×T1+W2×T2+W3×T3+ W4×T4; Wherein, T is the current loss value, W1 is the first weight coefficient, W2 is the second weight coefficient, W3 is the third weight coefficient, W4 is the fourth weight coefficient, and the sum of W1, W2, W3, and W4 is 1; The T1 is the ratio of the first cross entropy loss value to the preset first loss value, the T2 is the ratio of the second cross entropy loss value to the preset second loss value, the T3 is the ratio of the third cross entropy loss value to the preset third loss value, and the T4 is the ratio of the fourth cross entropy loss value to the preset fourth loss value.
[0008] It should be noted that the direct purpose of this application is to improve the efficiency of obtaining current radiology reports. Since there is no need for manual writing, a lot of human resources and time resources are saved, so the efficiency of obtaining current radiology reports can be improved, thus achieving the direct purpose of this application.
[0009] In a possible implementation of the first aspect, obtaining, in a preset medical knowledge graph, a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description includes: Obtaining a calling instruction, and calling a query interface of the medical knowledge graph by executing the calling instruction; Through the query interface, the preset entity name, the preset entity description, and the preset treatment plan corresponding to the preset image description are obtained in the medical knowledge graph.
[0010] In a possible implementation manner of the first aspect, if the current loss value is less than a preset loss value, obtaining a current radiological image of the radiological device, inputting the current radiological image into the trained deep neural network, and obtaining a current radiology report output by the trained deep neural network based on the current radiological image, includes: Adding the current loss values corresponding to the training samples in the training set to generate a total loss value of the training set, and dividing the total loss value by the number of the training samples to generate a target loss value; Update the model parameters in the deep neural network based on the target loss value to obtain the updated deep neural network; Training the updated deep neural network based on the training set, when the target loss value is less than the preset loss value, stopping the training of the updated deep neural network and saving the trained deep neural network; Acquire the current radiological image of the radiological device, input the current radiological image into the trained deep neural network, and acquire a current radiological report output by the trained deep neural network based on the current radiological image.
[0011] In a possible implementation of the first aspect, if the current loss value is less than a preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: Acquire a display page corresponding to the current radiological image, and acquire a first window and a second window of the display page; The current radiological image is displayed in the first window, and the current radiological report is displayed in the second window. The first window and the second window are different windows.
[0012] In a possible implementation of the first aspect, if the current loss value is less than a preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: An evaluation page of the current radiology report is obtained, evaluation information input on the evaluation page is obtained, and the evaluation information is stored.
[0013] In a possible implementation of the first aspect, if the current loss value is less than a preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: A target file is acquired, and the current radiological image and the current radiology report are stored in the target file.
[0014] In a possible implementation of the first aspect, if the current loss value is less than a preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: Read the preset upload time and determine whether the current time is the upload time; If the current time is the upload time, the preset server is connected to upload the current radiological image and the current radiology report to the server.
[0015] In a second aspect, an embodiment of the present application provides a radiology report acquisition device, which is applied to an electronic device, wherein the electronic device stores a deep neural network, including: A first acquisition module is used to acquire a preset radiological image of a radiological device and acquire a preset image description of the preset radiological image; A second acquisition module is used to obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description in a preset medical knowledge graph; a merging module, configured to merge the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and to form a training sample with the data item and a preset radiology report corresponding to the preset radiological image; An input module, used for training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; A third acquisition module is used to obtain a first cross entropy loss value between the predicted image description and the preset image description, obtain a second cross entropy loss value between the predicted entity name and the preset entity name, obtain a third cross entropy loss value between the predicted entity description and the preset entity description, and obtain a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme through a preset cross entropy loss function; A generation module, configured to generate a model according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value, to generate a current loss value corresponding to the training sample; The fourth acquisition module is used to acquire a current radiological image of the radiological equipment if the current loss value is less than a preset loss value, input the current radiological image into the trained deep neural network, and acquire a current radiology report output by the trained deep neural network based on the current radiological image.
[0016] In a third aspect, an embodiment of the present application 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 radiology report acquisition method of the first aspect described above is implemented.
[0017] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for obtaining a radiology report according to the first aspect is implemented.
[0018] In a fifth aspect, an embodiment of the present application provides a computer program product. When the computer program product is run on an electronic device, the electronic device executes the radiology report acquisition method of the first aspect.
[0019] The beneficial effects of the embodiment of the present application lie in two aspects. On the one hand, if the current loss value is less than the preset loss value, the current radiological image of the radiological equipment is obtained, and the current radiological image is input into the trained deep neural network to obtain the current radiological report output by the trained deep neural network based on the current radiological image. Since no manual writing is required, a large amount of human resources and time resources are saved, which is conducive to improving the efficiency of obtaining the current radiological report. On the other hand, through the deep neural network, it will not be affected by human factors, which is conducive to improving the reliability of the current radiological report obtained. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0021] Figure 1 An application scenario diagram of the radiology report acquisition method provided in an embodiment of the present application; Figure 2 is a flow chart of a method for obtaining a radiology report provided in an embodiment of the present application; Figure 3 A flowchart of calling the query interface provided in an embodiment of the present application; Figure 4 A schematic block diagram of a radiology report acquisition device provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.
[0023] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.
[0024] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.
[0025] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0026] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.
[0027] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.
[0028] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.
[0029] In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the fact that they can be implemented by ordinary technicians in this field. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.
[0030] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.
[0031] The radiology report acquisition method provided in the embodiment of the present application can be applied to electronic devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc. The embodiment of the present application does not impose any restrictions on the specific type of electronic devices.
[0032] See also Figure 1 , Figure 1 The application scenario diagram of the radiology report acquisition method provided in the embodiment of the present application is described in detail as follows: The electronic device is connected to the storage system, and obtains a preset radiological image of the radiological device through the storage system, and obtains a preset image description of the preset radiological image.
[0033] After generating the radiological image, the radiological equipment will automatically send the preset radiological image to the storage system.
[0034] The preset radiological images may come from one file of the storage system or from multiple files of the storage system.
[0035] For the sake of explanation, the preset radiological images are taken from two files as an example, as follows: The two files include file 1 and file 2. The electronic device reads file 1 and file 2 respectively, and obtains preset radiological images of the radiological device from file 1 and file 2.
[0036] In an embodiment of the present application, the electronic device can read different files at the same time and obtain preset radiological images of the radiological device from the different files.
[0037] See also Figure 2 , Figure 2 It is a flow chart of a method for obtaining a radiology report provided in an embodiment of the present application. The method can be applied to an electronic device storing a deep neural network.
[0038] like Figure 2 As shown, the radiology report acquisition method provided in the embodiment of the present application includes the following steps, which are described in detail as follows: S201, obtaining a preset radiological image of a radiological device, and obtaining a preset image description of the preset radiological image; Among them, a preset radiological image of the radiological equipment is obtained, visual features of the preset radiological image are extracted, a visual description of the visual features is obtained, and the visual description is used as a preset image description of the preset radiological image.
[0039] For example, to obtain the preset radiological image of the radiological equipment, the pre-trained rest152 is used as the feature extraction model, and the visual features of the preset radiological image are extracted through the feature extraction model.
[0040] Among them, ResNet152 is a type of deep residual network, which will not be described here.
[0041] Among them, visual features refer to information that can be automatically extracted from visual data such as images or videos for analysis or recognition. These features can be low-level, such as color, texture, shape, edge, etc., or high-level, such as the category of objects, the structure of scenes, etc.
[0042] Visual description refers to the process of summarizing, illustrating or explaining visual content in natural language. Descriptions can be simple names of objects in an image or complex narratives about the overall content of the image. In natural language processing, visual descriptions are usually represented as text sequences, which can contain words, phrases, sentences or longer paragraphs.
[0043] S202, obtaining a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description in a preset medical knowledge graph; The preset image description is a preset image description.
[0044] The preset entity name is a preset entity name.
[0045] The preset entity description is a preset entity description.
[0046] Among them, the preset processing solution is a preset processing solution.
[0047] For the sake of illustration, an example is given below: Image Description 1: There is a shadow.
[0048] Image Description 1: No shadow.
[0049] For the sake of illustration, an example is given below: Entity names are the basis of medical knowledge graphs and represent specific medical concepts or objects. Here are some examples of entity names: Inspection items: such as item 1, item 2, item 3, etc.
[0050] Department: Department 1, Department 2, Department 3, etc.
[0051] For the sake of illustration, an example is given below: The entity description is a further explanation and description of the entity name, including detailed information and characteristics of the entity. The following are some examples of entity descriptions: Electrocardiogram: A method of recording the electrical activity of the heart.
[0052] Department 1: A department that specializes in studying a certain subject.
[0053] For the sake of illustration, an example is given below: In the medical knowledge graph, a treatment plan usually refers to an examination or management plan. Here are some examples of treatment plans: Option 1: Diet adjustment.
[0054] Option 2: Regular check-ups.
[0055] S203, combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into one training sample; The feature vector of the data item and the feature vector of the preset radiology report corresponding to the preset radiological image are combined into a training sample.
[0056] S204, training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; When building a deep neural network, a pre-trained large language model can be used as an initialization parameter. This approach can accelerate the training process of the deep neural network and may improve the performance of the final deep neural network. Because the pre-trained large language model has learned a wealth of language knowledge and rules from a large-scale data set, this knowledge and rules can serve as a good starting point to help the deep neural network adapt to specific tasks more quickly.
[0057] Wherein, if the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, including: Adding the current loss values corresponding to the training samples in the training set to generate a total loss value of the training set, and dividing the total loss value by the number of the training samples to generate a target loss value; Update the model parameters in the deep neural network based on the target loss value to obtain the updated deep neural network; Training the updated deep neural network based on the training set, when the target loss value is less than the preset loss value, stopping the training of the updated deep neural network and saving the trained deep neural network; Acquire the current radiological image of the radiological device, input the current radiological image into the trained deep neural network, and acquire a current radiological report output by the trained deep neural network based on the current radiological image.
[0058] S205, obtaining, through a preset cross entropy loss function, a first cross entropy loss value between the predicted image description and the preset image description, obtaining a second cross entropy loss value between the predicted entity name and the preset entity name, obtaining a third cross entropy loss value between the predicted entity description and the preset entity description, and obtaining a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme; S206, generating a current loss value corresponding to the training sample according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value generation model; Exemplarily, the current loss value generation model is: T=W1×T1+W2×T2+W3×T3+ W4×T4; Wherein, T is the current loss value, W1 is the first weight coefficient, W2 is the second weight coefficient, W3 is the third weight coefficient, W4 is the fourth weight coefficient, and the sum of W1, W2, W3, and W4 is 1; The T1 is the ratio of the first cross entropy loss value to the preset first loss value, the T2 is the ratio of the second cross entropy loss value to the preset second loss value, the T3 is the ratio of the third cross entropy loss value to the preset third loss value, and the T4 is the ratio of the fourth cross entropy loss value to the preset fourth loss value.
[0059] Among them, the current loss value generation model is a generation model of the current loss value.
[0060] In this way, through the values of T1, T2, T3, and T4, we can know the degree of loss of the deep neural network in the four aspects of preset image description, preset entity name, preset entity description, and preset processing solution.
[0061] For the sake of illustration, an example is given below: Exemplarily, the T1 is greater than 1, indicating that the current difference between the predicted image description and the preset image description is greater than the expected difference between the predicted image description and the preset image description. The T1 is equal to 1, indicating that the current difference between the predicted image description and the preset image description is equal to the expected difference between the predicted image description and the preset image description; The T1 is less than 1, indicating that the current difference between the predicted image description and the preset image description is less than the expected difference between the predicted image description and the preset image description.
[0062] Exemplarily, the T2 is greater than 1, indicating that the current difference between the predicted entity name and the preset entity name is greater than the expected difference between the predicted entity name and the preset entity name. The T2 is equal to 1, indicating that the current difference between the predicted entity name and the preset entity name is equal to the expected difference between the predicted entity name and the preset entity name; The T2 is less than 1, indicating that the current difference between the predicted entity name and the preset entity name is less than the expected difference between the predicted entity name and the preset entity name.
[0063] Exemplarily, the T3 is greater than 1, indicating that the current difference between the predicted entity description and the preset entity description is greater than the expected difference between the predicted entity description and the preset entity description. The T3 is equal to 1, indicating that the current difference between the predicted entity description and the preset entity description is equal to the expected difference between the predicted entity description and the preset entity description; The T3 is less than 1, indicating that the current difference between the predicted entity description and the preset entity description is less than the expected difference between the predicted entity description and the preset entity description.
[0064] Exemplarily, T4 is greater than 1, indicating that the current difference between the predicted processing solution and the preset processing solution is greater than the expected difference between the predicted processing solution and the preset processing solution. The T4 is equal to 1, indicating that the current difference between the predicted treatment plan and the preset treatment plan is equal to the expected difference between the predicted treatment plan and the preset treatment plan; The T4 is less than 1, indicating that the current difference between the predicted processing solution and the preset processing solution is less than the expected difference between the predicted processing solution and the preset processing solution.
[0065] Among them, during the training process, by increasing W1, W2, W3, and W4, the learning effect of the deep neural network in the four aspects of preset image description, preset entity name, preset entity description, and preset processing solution can be improved.
[0066] For the sake of illustration, an example is given below: During the training process, if the deep neural network does not achieve ideal learning results for the preset image description, W1 can be adjusted and increased. By amplifying the contribution of T1 in the current loss value, the deep neural network can pay more attention to the preset image description, thereby improving the learning effect of the deep neural network for the preset image description.
[0067] During the training process, if the deep neural network does not learn the preset entity names well, W2 can be adjusted and increased. By amplifying the contribution of T2 in the current loss value, the deep neural network can increase its attention to the preset entity names, thereby improving the learning effect of the deep neural network on the preset entity names.
[0068] During the training process, if the deep neural network does not learn the preset entity description well, W3 can be adjusted and increased. By amplifying the contribution of T3 in the current loss value, the deep neural network can increase its attention to the preset entity description, thereby improving the learning effect of the deep neural network for the preset entity description.
[0069] During the training process, if the deep neural network does not learn the preset processing solution well, W4 can be adjusted and increased. By amplifying the contribution of T4 in the current loss value, the deep neural network can increase its attention to the preset processing solution, thereby improving the learning effect of the deep neural network for the preset processing solution.
[0070] Exemplarily, the higher the current loss value, the larger the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, and the fourth cross entropy loss value; the lower the current loss value, the smaller the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, and the fourth cross entropy loss value.
[0071] S207, if the current loss value is less than the preset loss value, obtain the current radiological image of the radiological equipment, input the current radiological image into the trained deep neural network, and obtain the current radiological report output by the trained deep neural network based on the current radiological image.
[0072] Among them, W1, W2, W3, and W4 are set by the user or by system default, and are not limited here.
[0073] For the sake of illustration, an example is given below: For example, W1 is 0.25, W2 is 0.25, W3 is 0.25, W4 is 0.25, T1 is 1.2, T2 is 1.1, T3 is 1.3, and T4 is 1; T=0.25×1.2+0.25×1.1+0.25×1.3+0.25×1.3=1.22; For example, W1 is 0.3, W2 is 0.3, W3 is 0.2, W4 is 0.2, T1 is 1.3, T2 is 1.1, T3 is 1.3, and T3 is 1.3; T=0.3×1.3+0.3×1.1+0.2×1.3+0. 2×1.3=1.24; The preset loss value is set by the user or by the system default and is not restricted here.
[0074] For the sake of illustration, an example is given below: For example, the default loss value is 1.1; When the current loss value is 1.13, the current loss value is less than the preset loss value. Therefore, the electronic device obtains the current radiological image of the radiation device, inputs the current radiological image into the trained deep neural network, and obtains the current radiology report output by the trained deep neural network based on the current radiological image.
[0075] For example, the default loss value is 1.12; When the current loss value is 1.13, the current loss value is less than the preset loss value. Therefore, the electronic device obtains the current radiological image of the radiation device, inputs the current radiological image into the trained deep neural network, and obtains the current radiology report output by the trained deep neural network based on the current radiological image.
[0076] Among them, when the current loss value is less than the preset loss value, it means that the current loss value is better than expected, which is conducive to ensuring the reliability of the deep neural network after training.
[0077] It should be noted that the direct purpose of this application is to improve the efficiency of obtaining current radiology reports. Since there is no need for manual writing, a lot of human resources and time resources are saved, so the efficiency of obtaining current radiology reports can be improved, thus achieving the direct purpose of this application.
[0078] Wherein, if the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained, and the radiological report acquisition method includes: Step A, obtaining a display page corresponding to the current radiological image, and obtaining a first window and a second window of the display page; The current radiological image is displayed in the first window, and the current radiological report is displayed in the second window. The first window and the second window are different windows.
[0079] Wherein, if the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained, and the radiological report acquisition method includes: Step B, obtaining the evaluation page of the current radiology report, obtaining the evaluation information input on the evaluation page, and storing the evaluation information.
[0080] Among them, the deep neural network can be further optimized through evaluation information.
[0081] Wherein, if the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained, and the radiological report acquisition method includes: Step C, obtaining a target file, and storing the current radiological image and the current radiology report in the target file.
[0082] Wherein, if the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained, and the radiological report acquisition method includes: Step D, reading the preset upload time, and determining whether the current time is the upload time; If the current time is the upload time, the preset server is connected to upload the current radiological image and the current radiology report to the server.
[0083] Step A can be performed before or after step B, step C, or step D; Step B can be performed before or after step A, step C, or step D; Step C can be performed before or after step A, step B, or step D; Step D can be performed before or after step A, step B, or step C; Step A may also be executed simultaneously with step B, step C, and step D. The specific execution order is not limited here.
[0084] The beneficial effects of the embodiment of the present application lie in two aspects. On the one hand, if the current loss value is less than the preset loss value, the current radiological image of the radiological equipment is obtained, and the current radiological image is input into the trained deep neural network to obtain the current radiological report output by the trained deep neural network based on the current radiological image. Since no manual writing is required, a large amount of human resources and time resources are saved, which is conducive to improving the efficiency of obtaining the current radiological report. On the other hand, through the deep neural network, it will not be affected by human factors, which is conducive to improving the reliability of the current radiological report obtained.
[0085] See also Figure 3 , Figure 3 The flowchart of calling the query interface provided in the embodiment of the present application is detailed as follows: S301, obtaining a call instruction, and calling a query interface of the medical knowledge graph by executing the call instruction; S302, through the query interface, in the medical knowledge graph, obtain the preset entity name, the preset entity description, and the preset treatment plan corresponding to the preset image description.
[0086] In the embodiment of the present application, the efficiency of obtaining the preset entity name, the preset entity description, and the preset processing solution corresponding to the preset image description can be improved through the query interface.
[0087] For the radiology report acquisition method described in the above embodiment, please refer to Figure 4 , Figure 4 A schematic block diagram of a radiology report acquisition device provided in an embodiment of the present application, Figure 4 The radiology report acquisition device 400 shown can be applied to Figure 1 The electronic device in the application scenario diagram shown below takes the electronic device as an example. Figure 4 The radiology report acquisition device 400 shown is described in detail. The radiology report acquisition device 400 may include a first acquisition module 401, a second acquisition module 402, a merging module 403, an input module 404, a third acquisition module 405, a generation module 406, and a fourth acquisition module 407.
[0088] The first acquisition module 401 is used to acquire a preset radiological image of a radiological device and acquire a preset image description of the preset radiological image; The second acquisition module 402 is used to obtain the preset entity name, preset entity description, and preset treatment plan corresponding to the preset image description in the preset medical knowledge graph; A merging module 403 is used to merge the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and to form a training sample with the data item and a preset radiology report corresponding to the preset radiological image; An input module 404 is used to train the deep neural network based on the training samples, input the data items of the training samples into the deep neural network, obtain a predicted radiology report output by the deep neural network based on the data items, and obtain a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; The third acquisition module 405 is used to obtain a first cross entropy loss value between the predicted image description and the preset image description, obtain a second cross entropy loss value between the predicted entity name and the preset entity name, obtain a third cross entropy loss value between the predicted entity description and the preset entity description, and obtain a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme through a preset cross entropy loss function; A generating module 406, configured to generate a model according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value, to generate a current loss value corresponding to the training sample; The fourth acquisition module 407 is used to acquire the current radiological image of the radiological device if the current loss value is less than the preset loss value, input the current radiological image into the trained deep neural network, and acquire the current radiological report output by the trained deep neural network based on the current radiological image.
[0089] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0090] The beneficial effects of the embodiment of the present application lie in two aspects. On the one hand, if the current loss value is less than the preset loss value, the current radiological image of the radiological equipment is obtained, and the current radiological image is input into the trained deep neural network to obtain the current radiological report output by the trained deep neural network based on the current radiological image. Since no manual writing is required, a large amount of human resources and time resources are saved, which is conducive to improving the efficiency of obtaining the current radiological report. On the other hand, through the deep neural network, it will not be affected by human factors, which is conducive to improving the reliability of the current radiological report obtained.
[0091] See also Figure 5 , Figure 5A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.
[0092] like Figure 5 As shown, Figure 5 The electronic device 2 includes: at least one processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the at least one processor 20, and when the processor 20 executes the computer program 22, the steps in any of the above-mentioned method embodiments are implemented.
[0093] The electronic device 2 may include, but is not limited to, a processor 20 and a memory 21. Those skilled in the art will appreciate that Figure 5 It is only an example of the electronic device 2 and does not constitute a limitation on the electronic device 2. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.
[0094] The processor 20 is used to run the computer program 22 stored in the memory 21, and implement the following steps when executing the computer program 22: Acquire a preset radiographic image of a radiological device, and acquire a preset image description of the preset radiographic image; In a preset medical knowledge graph, obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description; Combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into a training sample; Training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; Through a preset cross-entropy loss function, a first cross-entropy loss value between the predicted image description and the preset image description is obtained, a second cross-entropy loss value between the predicted entity name and the preset entity name is obtained, a third cross-entropy loss value between the predicted entity description and the preset entity description is obtained, and a fourth cross-entropy loss value between the predicted processing scheme and the preset processing scheme is obtained; a model is generated according to the first cross-entropy loss value, the second cross-entropy loss value, the third cross-entropy loss value, the fourth cross-entropy loss value and a preset current loss value to generate a current loss value corresponding to the training sample; if the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiology report output by the trained deep neural network based on the current radiological image is obtained.
[0095] In some embodiments, the processor 20 is configured to implement: Obtaining a calling instruction, and calling a query interface of the medical knowledge graph by executing the calling instruction; Through the query interface, the preset entity name, the preset entity description, and the preset treatment plan corresponding to the preset image description are obtained in the medical knowledge graph.
[0096] In some embodiments, the processor 20 is configured to implement: Adding the current loss values corresponding to the training samples in the training set to generate a total loss value of the training set, and dividing the total loss value by the number of the training samples to generate a target loss value; Update the model parameters in the deep neural network based on the target loss value to obtain the updated deep neural network; Training the updated deep neural network based on the training set, when the target loss value is less than the preset loss value, stopping the training of the updated deep neural network and saving the trained deep neural network; Acquire the current radiological image of the radiological device, input the current radiological image into the trained deep neural network, and acquire a current radiological report output by the trained deep neural network based on the current radiological image.
[0097] In some embodiments, the processor 20 is configured to implement: Acquire a display page corresponding to the current radiological image, and acquire a first window and a second window of the display page; The current radiological image is displayed in the first window, and the current radiological report is displayed in the second window. The first window and the second window are different windows.
[0098] In some embodiments, the processor 20 is configured to implement: An evaluation page of the current radiology report is obtained, evaluation information input on the evaluation page is obtained, and the evaluation information is stored.
[0099] In some embodiments, the processor 20 is configured to implement: A target file is acquired, and the current radiological image and the current radiology report are stored in the target file.
[0100] In some embodiments, the processor 20 is configured to implement: Read the preset upload time and determine whether the current time is the upload time; If the current time is the upload time, the preset server is connected to upload the current radiological image and the current radiology report to the server.
[0101] The processor 20 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0102] In some embodiments, the memory 21 may be an internal storage unit of the electronic device 2, such as a hard disk or memory of the electronic device 2. In other embodiments, the memory 21 may also be an external storage device of the electronic device 2, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 2. Further, the memory 21 may also include both an internal storage unit and an external storage device of the electronic device 2. The memory 21 is used to store an operating system, an application program, a boot loader, data, and other programs, such as the program code of the computer program. The memory 21 may also be used to temporarily store data that has been output or is to be output.
[0103] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.
[0104] An embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.
[0105] The computer-readable storage medium stores program codes, which can be called by a processor to execute the radiology report acquisition method described in the above method embodiment.
[0106] The computer-readable storage medium has a storage space for program codes.
[0107] The program code includes the code of any step in the radiology report acquisition method described in the above method embodiment.
[0108] For example, the program code is called by the processor and can execute the following steps: Acquire a preset radiographic image of a radiological device, and acquire a preset image description of the preset radiographic image; In a preset medical knowledge graph, obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description; Combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into a training sample; Training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; By using a preset cross entropy loss function, a first cross entropy loss value between the predicted image description and the preset image description is obtained, a second cross entropy loss value between the predicted entity name and the preset entity name is obtained, a third cross entropy loss value between the predicted entity description and the preset entity description is obtained, and a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme is obtained; Generate a current loss value corresponding to the training sample according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value generation model; If the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained.
[0109] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.
[0110] Among them, the computer-readable storage medium can also be an external storage device of the radiology report acquisition device or electronic device, for example, a plug-in hard disk equipped on the radiology report acquisition device or electronic device, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), a non-volatile computer-readable medium (non-transitory computer-readable storage medium), etc.
[0111] Since the computer program stored in the computer-readable storage medium can execute any one of the radiology report acquisition methods provided in the embodiments of the present application, the computer-readable storage medium can achieve the beneficial effects that can be achieved by any one of the radiology report acquisition methods provided in the embodiments of the present application. Please see the previous embodiments for details and will not be repeated here.
[0112] An embodiment of the present application provides a computer program product. When the computer program product is executed on an electronic device, the electronic device executes the above-mentioned radiology report acquisition method.
[0113] The computer program product is loaded into the electronic device and can execute the following steps: Acquire a preset radiographic image of a radiological device, and acquire a preset image description of the preset radiographic image; In a preset medical knowledge graph, obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description; Combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into a training sample; Training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; By using a preset cross entropy loss function, a first cross entropy loss value between the predicted image description and the preset image description is obtained, a second cross entropy loss value between the predicted entity name and the preset entity name is obtained, a third cross entropy loss value between the predicted entity description and the preset entity description is obtained, and a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme is obtained; Generate a current loss value corresponding to the training sample according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value generation model; If the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained.
[0114] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
[0115] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.
[0116] Based on such understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electrical carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.
[0117] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0118] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.
Claims
1. A method for obtaining a radiology report, characterized in that: Applied to an electronic device, the electronic device stores a deep neural network, and the radiology report acquisition method includes: Acquire a preset radiographic image of a radiological device, and acquire a preset image description of the preset radiographic image; In a preset medical knowledge graph, obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description; Combining the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and combining the data item and the preset radiology report corresponding to the preset radiological image into one training sample; Training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; By using a preset cross entropy loss function, a first cross entropy loss value between the predicted image description and the preset image description is obtained, a second cross entropy loss value between the predicted entity name and the preset entity name is obtained, a third cross entropy loss value between the predicted entity description and the preset entity description is obtained, and a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme is obtained; Generate a current loss value corresponding to the training sample according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value generation model; If the current loss value is less than the preset loss value, the current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and the current radiological report output by the trained deep neural network based on the current radiological image is obtained.
2. The method for obtaining radiology reports according to claim 1, characterized in that: The step of obtaining, in the preset medical knowledge graph, a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description includes: Obtaining a calling instruction, and calling a query interface of the medical knowledge graph by executing the calling instruction; Through the query interface, the preset entity name, the preset entity description, and the preset treatment plan corresponding to the preset image description are obtained in the medical knowledge graph.
3. The method for obtaining radiology reports according to claim 1, characterized in that: If the current loss value is less than a preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiology report output by the trained deep neural network based on the current radiological image is obtained, including: Adding the current loss values corresponding to the training samples in the training set to generate a total loss value of the training set, and dividing the total loss value by the number of the training samples to generate a target loss value; Update the model parameters in the deep neural network based on the target loss value to obtain the updated deep neural network; Training the updated deep neural network based on the training set, when the target loss value is less than the preset loss value, stopping the training of the updated deep neural network and saving the trained deep neural network; Acquire the current radiological image of the radiological device, input the current radiological image into the trained deep neural network, and acquire a current radiological report output by the trained deep neural network based on the current radiological image.
4. The method for obtaining radiology reports according to claim 1, characterized in that: If the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: Acquire a display page corresponding to the current radiological image, and acquire a first window and a second window of the display page; The current radiological image is displayed in the first window, and the current radiological report is displayed in the second window. The first window and the second window are different windows.
5. The method for obtaining radiology reports according to claim 1, characterized in that: If the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: An evaluation page of the current radiology report is obtained, evaluation information input on the evaluation page is obtained, and the evaluation information is stored.
6. The method for obtaining radiology reports according to claim 1, characterized in that: If the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: A target file is acquired, and the current radiological image and the current radiology report are stored in the target file.
7. The method for obtaining radiology reports according to claim 1, characterized in that: If the current loss value is less than the preset loss value, a current radiological image of the radiological device is obtained, the current radiological image is input into the trained deep neural network, and a current radiological report output by the trained deep neural network based on the current radiological image is obtained, the radiological report obtaining method includes: Read the preset upload time and determine whether the current time is the upload time; If the current time is the upload time, the preset server is connected to upload the current radiological image and the current radiology report to the server.
8. A radiology report acquisition device, characterized in that: Applied to an electronic device, the electronic device storing a deep neural network, comprising: A first acquisition module is used to acquire a preset radiological image of a radiological device and acquire a preset image description of the preset radiological image; A second acquisition module is used to obtain a preset entity name, a preset entity description, and a preset treatment plan corresponding to the preset image description in a preset medical knowledge graph; a merging module, configured to merge the preset image description, the preset entity name, the preset entity description, and the preset treatment plan into one data item, and to form a training sample with the data item and a preset radiology report corresponding to the preset radiological image; An input module, used for training the deep neural network based on the training samples, inputting the data items of the training samples into the deep neural network, obtaining a predicted radiology report output by the deep neural network based on the data items, and obtaining a predicted image description, a predicted entity name, a predicted entity description, and a predicted treatment plan in the predicted radiology report; A third acquisition module is used to obtain a first cross entropy loss value between the predicted image description and the preset image description, obtain a second cross entropy loss value between the predicted entity name and the preset entity name, obtain a third cross entropy loss value between the predicted entity description and the preset entity description, and obtain a fourth cross entropy loss value between the predicted processing scheme and the preset processing scheme through a preset cross entropy loss function; A generation module, configured to generate a model according to the first cross entropy loss value, the second cross entropy loss value, the third cross entropy loss value, the fourth cross entropy loss value and a preset current loss value, to generate a current loss value corresponding to the training sample; The fourth acquisition module is used to acquire the current radiological image of the radiological equipment if the current loss value is less than the preset loss value, input the current radiological image into the trained deep neural network, and acquire the current radiological report output by the trained deep neural network based on the current radiological image.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for obtaining a radiology report according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for obtaining a radiology report according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Image report generation method and device, computer equipment and storage medium
CN112530550A
Large language model training method and text emotion entity recognition method
CN118277569A
Medical image report quality control method, device, equipment and medium
CN118471423A
Medical image report intelligent generation method and device, equipment and storage medium
CN118506956A
