Pathological report labeling and model training method, system, equipment and medium
Through multiple iterations, the accuracy of pathological report labeling is improved, the subjectivity and uncertainty of the labeling results in the existing technology are solved, and the performance of the deep learning model is improved.
Patent Information
- Application Number
- CN202411221499.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-27
AI Technical Summary
The existing pathological report labeling methods only rely on keywords manually screened once, resulting in subjectivity and uncertainty in the annotation results, affecting the training performance of deep learning models.
By obtaining multiple pathological reports, extracting high-frequency words and performing multiple iterative screening, a keyword list for the target disease is constructed, and the pathological reports to be marked are used to improve the accuracy of the label.
Improve the accuracy of pathological report labeling, enhance the performance of deep learning models, and reduce the impact of labeling inaccuracy.
Smart Images
Figure CN119207693A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the medical field, and in particular to pathology report annotation and model training methods, systems, equipment, and media. Background Art
[0002] In the medical field, especially in pathology and imaging diagnosis, accurately judging whether a nodule is benign or malignant is an important part of clinical diagnosis and treatment. At present, the judgment of whether a nodule is benign or malignant mainly depends on the doctor's experience and the results of imaging examinations.
[0003] With the development of machine learning, deep learning models have been widely used in medical image analysis due to their excellent performance in detection and classification tasks. Deep learning models learn complex feature representations through a large amount of pathological report data corresponding to breast nodules, thereby achieving high-precision nodule detection and benign and malignant classification. However, high-quality annotated datasets are essential for training accurate and reliable deep learning models.
[0004] At present, most studies have manually screened the keywords in the pathology reports once, and then directly annotated the pathology reports based on the keywords after the manual screening. However, the interpretation process of the pathology report is relatively complicated. The results of annotating the pathology reports based on the keywords in the pathology reports only after manual screening are often subjective and uncertain, which may affect the accuracy of the pathology report annotation, resulting in poor performance of the model trained by the deep learning model using the inaccurately annotated pathology reports.
[0005] Therefore, there is an urgent need for a method that can more accurately annotate pathology reports to improve the accuracy of the diagnostic results corresponding to the pathology reports. Summary of the invention
[0006] The present application provides a pathology report annotation and model training method, system, device, and medium to solve the technical problem that the existing pathology report to be annotated is annotated based only on a manual screening of keywords in the pathology report, which may lead to inaccurate annotated pathology report results.
[0007] The first aspect of the present application provides a pathology report annotation method, the method comprising:
[0008] Obtain multiple pathology reports corresponding to the target disease in a first preset year;
[0009] Extracting a plurality of high-frequency words from a plurality of the pathology reports corresponding to the target disease according to a preset standard, and forming a high-frequency word list with the plurality of the high-frequency words;
[0010] Acquire multiple keywords in the high-frequency word list and form a keyword list, wherein the keywords are obtained by manually screening multiple high-frequency words in the high-frequency word list;
[0011] Performing multiple iterative screening on all the keywords in the keyword table, and constructing a target keyword table corresponding to the target disease based on the results of the multiple iterative screening;
[0012] Based on the target keyword table corresponding to the target disease, the acquired pathology report corresponding to the target disease to be labeled is labeled.
[0013] In some embodiments of the present application, the target keyword table corresponding to the target disease is constructed in the following manner:
[0014] Verifying all the keywords in the keyword table word by word using a plurality of pathology reports corresponding to the target disease in the first preset year, and constructing a verification word table corresponding to the target disease based on the verified results;
[0015] All the keywords in the verification vocabulary are tested word by word using a plurality of pathology reports corresponding to the target disease in a second preset year, and a target keyword table corresponding to the target disease is constructed based on the test results.
[0016] In some embodiments of the present application, the verification vocabulary corresponding to the target disease is constructed in the following manner:
[0017] All keywords in the keyword table are verified word by word using multiple pathology reports corresponding to the target disease in the first preset year, and the verified results are manually streamlined, and a verification word table corresponding to the target disease is constructed based on the results after manual streamlining.
[0018] In some embodiments of the present application, the acquired pathology report corresponding to the target disease to be annotated is annotated in the following manner:
[0019] All keywords in the target keyword table corresponding to the target disease are graded according to a preset grade allocation rule, and based on the target keyword table corresponding to the target disease after the grade setting, the acquired pathology report corresponding to the target disease to be marked is marked.
[0020] In some embodiments of the present application, the target disease is breast nodules.
[0021] A second aspect of the present application provides a pathology report annotation system, the system comprising:
[0022] A data acquisition module, used to acquire multiple pathology reports corresponding to the target disease in a first preset year;
[0023] A high-frequency word extraction module, used to extract a plurality of high-frequency words from a plurality of the pathology reports corresponding to the target disease according to a preset standard, and form a high-frequency word list with the plurality of the high-frequency words;
[0024] A keyword extraction module, used for obtaining a plurality of keywords in the high-frequency word list and forming a keyword list, wherein the keywords are obtained by manually screening a plurality of high-frequency words in the high-frequency word list;
[0025] A multiple quality control module, used for performing multiple iterative screening on all the keywords in the keyword table, and constructing a target keyword table corresponding to the target disease based on the results of multiple iterative screening;
[0026] The marking module is used to mark the acquired pathology report corresponding to the target disease to be marked based on the target keyword table corresponding to the target disease.
[0027] The third aspect of the present application provides a method for training a disease detection model, the method comprising:
[0028] Obtain multiple pathology reports corresponding to breast nodules;
[0029] Annotating the multiple pathology reports of the breast nodules using the pathology report annotation method described in any of the above embodiments;
[0030] A training set is constructed based on the multiple pathological reports of the labeled breast nodules, and the breast nodule detection model is iteratively trained multiple times using the training set until convergence.
[0031] A fourth aspect of the present application provides an application method of a disease detection model, the method comprising:
[0032] Obtained pathology reports for testing;
[0033] The disease detection model trained by the disease detection model training method described in the above embodiment is used to obtain the detection result based on the pathology report to be detected.
[0034] The fifth aspect 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 the processor implements the steps of the method described in any one of the first aspect, the third aspect, and the fourth aspect of the above-mentioned embodiments when executing the computer program.
[0035] The sixth aspect of the present application provides a computer-readable storage medium having program instructions stored thereon, and when the program instructions are executed by a processor, the method described in any one of the first aspect, the third aspect, and the fourth aspect of the above-mentioned embodiments is implemented.
[0036] The present application has the following beneficial effects: the pathology report annotation scheme of the present application first extracts high-frequency words in the pathology report corresponding to the target disease, then manually selects keywords from the high-frequency words, and then iteratively selects the selected keywords multiple times to obtain keywords corresponding to the target disease, and finally uses the keywords after multiple selections to annotate the pathology report corresponding to the target disease to be annotated. Compared with the existing method of annotating the pathology report to be annotated based on only one manual screening of keywords in the pathology report, the pathology report annotation method of the present application improves the accuracy of pathology report annotation by multiple verifications of the results after manual screening, thereby using the accurately annotated pathology reports to train the deep learning model to improve the performance of the model. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and are used together with the specification to illustrate the technical solution of the present application.
[0038] Figure 1 It is a flowchart of an embodiment of a pathology report annotation method provided by the present application;
[0039] Figure 2 It is a flow chart of an embodiment of a method for constructing a target keyword table provided by the present application;
[0040] Figure 3 It is a schematic diagram of the framework of an embodiment of a pathology report annotation system provided by the present application;
[0041] Figure 4 This is an example flow chart that uses the pathology reports of the target hospital in the past two years to illustrate the pathology report annotation scheme of this application;
[0042] Figure 5 It is a flowchart of an embodiment of a disease detection model training method provided by the present application;
[0043] Figure 6 It is a flowchart of an embodiment of a disease detection model application method provided by the present application;
[0044] Figure 7 It is a schematic diagram of a framework of an embodiment of an electronic device provided by the present application;
[0045] Figure 8 It is a schematic diagram of a framework of an embodiment of a computer-readable storage medium provided by the present application. DETAILED DESCRIPTION
[0046] The scheme of the embodiment of the present application is described in detail below in conjunction with the drawings of the specification.
[0047] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.
[0048] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the objects associated before and after are in an "or" relationship. In addition, "many" in this article means two or more than two. In addition, the term "at least one" in this article means any combination of at least two of any one or more of a plurality of, for example, including at least one of A, B, and C, can mean including any one or more elements selected from the set consisting of A, B, and C.
[0049] As described in the background technology, the existing pathology reports are annotated based only on manual screening of keywords in the pathology reports, and the results are often subjective and uncertain, which may affect the accuracy of the pathology report annotations, thereby making the model performance of the deep learning model trained with inaccurately annotated pathology reports poor.
[0050] In order to solve the above problems, the present application provides a new pathology report annotation scheme, which extracts high-frequency words in the pathology report corresponding to the target disease, and then uses manual screening to select keywords from the high-frequency words, and then iterates and screens the selected keywords multiple times to obtain the keywords corresponding to the target disease, and finally uses the keywords after multiple screening to annotate the pathology report corresponding to the target disease to be annotated. The pathology report annotation scheme of the present application is used to annotate the pathology report, which improves the accuracy of the pathology report annotation, so that the deep learning model is trained using the pathology report with more accurate annotation to improve the performance of the model.
[0051] The present application is described in detail below with reference to the accompanying drawings and embodiments.
[0052] According to one embodiment of the present application, Figure 1 As shown, the present application provides a pathology report annotation method, the method comprising: S1, obtaining a plurality of pathology reports corresponding to a target disease in a first preset year;
[0053] S2, extracting multiple high-frequency words from multiple pathology reports corresponding to the target disease according to preset standards, and forming a high-frequency word list with the multiple high-frequency words; S3, obtaining multiple keywords in the high-frequency word list and forming a keyword list, wherein the keywords are obtained by manually screening multiple high-frequency words in the high-frequency word list; S4, performing multiple iterative screening on all the keywords in the keyword list, and constructing a target keyword list corresponding to the target disease based on the results of multiple iterative screening;
[0054] S5: Based on the target keyword table corresponding to the target disease, annotate the acquired pathology report corresponding to the target disease to be annotated. Steps S2, S4 and S5 are described in detail below.
[0055] 1. Step S2
[0056] In step S2, multiple high-frequency words in the multiple pathology reports corresponding to the target disease are extracted according to preset standards, and the multiple high-frequency words are combined into a high-frequency word list. According to one embodiment of the present application, a preset extraction method is used to extract multiple high-frequency words in the multiple pathology reports corresponding to the target disease according to preset standards. In addition, according to another embodiment of the present application, the preset extraction method is a program extraction method or a manual extraction method. Schematically, the manual extraction method is to extract according to the experience of the physician.
[0057] It can be seen from the above embodiments that by using a preset extraction method (such as program extraction, manual extraction, etc.) to extract multiple high-frequency words in multiple pathology reports corresponding to the target disease, the high-frequency words in the pathology report can be accurately obtained, thereby helping to reduce the workload of keyword extraction in subsequent pathology reports.
[0058] 2. Step S4
[0059] In step S4, all keywords in the keyword table are screened multiple times, and a target keyword table corresponding to the target disease is constructed based on the results of the multiple times of screening. Figure 2 As shown, the target keyword table corresponding to the target disease is obtained by executing the following steps S41-S42.
[0060] In step 41, all the keywords in the keyword table are verified word by word using multiple pathology reports corresponding to the target disease in the first preset year, and a verification word table corresponding to the target disease is constructed based on the verified results. In addition, according to an embodiment of the present application, all the keywords in the keyword table are verified word by word using multiple pathology reports corresponding to the target disease in the first preset year, and the verified results are manually streamlined, and a verification word table corresponding to the target disease is constructed based on the results after the manual streamlining.
[0061] In the step S42, all the keywords in the verification vocabulary are tested word by word using a plurality of pathology reports corresponding to the target disease in the second preset year, and a target keyword table corresponding to the target disease is constructed based on the test results.
[0062] It can be seen from the above embodiments that the present application realizes effective screening of keywords in pathology reports and prevents the generation of ambiguous words by multiple verification and updating of keywords in the keyword table, thereby extracting the keywords that best match the pathology report and improving the accuracy of pathology report annotation.
[0063] Step S5
[0064] In the step S5, based on the target keyword table corresponding to the target disease, the acquired pathological report corresponding to the target disease to be marked is marked. According to one embodiment of the present application, all keywords in the target keyword table corresponding to the target disease are graded according to a preset grade allocation rule, and based on the target keyword table corresponding to the target disease after the grade setting, the acquired pathological report corresponding to the target disease to be marked is marked. For example, all keywords in the target keyword table are set to five levels of 4, 3, 2, 1, and -1 according to the degree of malignancy, where -1 is benign, and the remaining 1, 2, 3, and 4 represent the degree of malignancy, so that the malignant description is preferentially extracted when extracting keywords, and a certain disambiguation effect can also be achieved. It should be noted that the above setting of the keyword level to 5 is only an example. In order to improve the accuracy of pathological report annotation, the specific level setting can be determined according to the annotation situation or task characteristics, and is not specifically limited here.
[0065] It can be seen from the above embodiments that by setting the levels of all keywords in the target keyword table, pathology reports can be classified into the most appropriate categories. Using pathology reports of multiple categories that have been labeled to train a deep learning model can help improve the performance of the model.
[0066] Combine the following Figure 3 The pathology report annotation scheme of this application is introduced in detail from the system side.
[0067] According to one embodiment of the present application, Figure 3 As shown, the present application provides a pathology report annotation system, the system comprising: a data acquisition module, used to acquire multiple pathology reports corresponding to a target disease in a first preset year; a high-frequency word extraction module, used to extract multiple high-frequency words in the multiple pathology reports corresponding to the target disease according to preset standards, and form a high-frequency word list with the multiple high-frequency words; a keyword extraction module, used to acquire multiple keywords in the high-frequency word list and form them into a keyword list, wherein the keywords are obtained by manually screening the multiple high-frequency words in the high-frequency word list; a multiple quality control module, used to perform multiple iterative screening on all the keywords in the keyword list, and construct a target keyword list corresponding to the target disease based on the results of multiple iterative screening; a tagging module, used to tag the acquired pathology report corresponding to the target disease to be labeled based on the target keyword list corresponding to the target disease.
[0068] In order to more clearly understand the pathology report annotation scheme proposed in this application, all pathology reports of the target hospital in 2022 are taken as the pathology reports corresponding to the target disease in the first preset year, and all pathology reports of the target hospital in 2023 are taken as the pathology reports corresponding to the target disease in the second preset year. The pathology report annotation scheme of this application is explained below using the above two years of pathology reports.
[0069] like Figure 4 As shown, it shows an example of using the pathology reports of the target hospital in the past two years to illustrate the pathology report annotation scheme of this application. Figure 4It can be seen that, first, the program extracts high-frequency words that appear frequently from the pathology report in 2022 and forms a high-frequency word list; second, the high-frequency word list is screened by manual selection to select suitable keywords and form a keyword list; third, the pathology report in 2022 is used to verify each keyword in the keyword list word by word to obtain valid keywords after verification, wherein, after verifying the keywords in the keyword list using the pathology report in 2022, a matched keyword list and an unmatched keyword list are obtained; fifth, the matched keyword list and the unmatched keyword list are screened by manual simplification to obtain the optimized verification word list; Sixth, use the pathology report of 2023 to test each keyword in the verification vocabulary word by word to obtain the target keyword table, where the key point of the test is to remove ambiguity rather than increase the proportion of annotated pathology reports and reduce the proportion of unannotated pathology reports; Seventh, obtain pathology reports of other years, and annotate the pathology reports of other years by the annotation program to obtain annotated pathology reports and unannotated pathology reports corresponding to other years; Eighth, use the program to extract high-frequency words that appear frequently in the unannotated pathology reports corresponding to other years and form a high-frequency vocabulary, and iterate the first to seventh schemes to complete the annotation task of all pathology reports in other years. It should be noted that using pathology reports of 2022 and 2023 to implement the annotation task of pathology reports of other years is only an example. The pathology reports of the first preset year and the pathology reports of the second preset year can be adjusted according to specific circumstances in actual applications, and no specific limitation is made here.
[0070] In summary, compared with the existing method of annotating the pathology report to be annotated based only on the keywords in the pathology report manually screened once, the pathology report annotation method of the present application improves the accuracy of pathology report annotation by multiple verifications of the results after manual screening, thereby using accurately annotated pathology reports to train the deep learning model to improve the performance of the model.
[0071] In addition, by performing a puncture biopsy or surgical resection on the nodule, a tissue sample is obtained, and a pathological analysis is performed to obtain a pathological report. The pathological report is the gold standard for judging whether the nodule is benign or malignant. To this end, according to one embodiment of the present application, the target disease is a breast nodule. It should be noted that the pathological report annotation method of the present application can be used to annotate pathological reports of other diseases, and breast nodules are only one of the embodiments, which is not specifically limited here.
[0072] In the medical field, especially in pathology and imaging diagnosis, accurately judging the benign and malignant nature of breast nodules is an important part of clinical diagnosis and treatment. At present, the judgment of the benign and malignant nature of breast nodules mainly depends on the doctor's experience and the results of imaging examinations. However, the process of manual judgment is often subjective and uncertain, which may lead to the inaccuracy of the diagnosis results. To this end, after annotating the pathology report corresponding to the breast nodule using the pathology report annotation scheme of the above embodiment of the present application, a deep learning model is trained using the annotated pathology report as a data set to assist doctors in making more accurate diagnostic results based on the pathology report. According to one embodiment of the present application, Figure 5 As shown, a training method for a disease detection model is provided, the method comprising: T1, obtaining multiple pathology reports corresponding to breast nodules; T2, annotating the multiple pathology reports of the breast nodules using the pathology report annotation scheme in any of the above embodiments; T3, constructing a training set based on the annotated multiple pathology reports of the breast nodules, and using the training set to perform multiple iterative training on the breast nodule detection model until convergence.
[0073] It can be seen from the above embodiments that using more accurately labeled pathology reports to train deep learning models can improve the performance of the model, thereby improving the accuracy of the diagnosis results of the pathology reports.
[0074] In addition, according to one embodiment of the present application, Figure 6 As shown, the present application provides an application method of a disease detection model, the method comprising: P1, obtaining a pathology report to be detected; P2, using the disease detection model training method described in the above embodiment to obtain a disease detection model, and obtaining a detection result based on the pathology report to be detected.
[0075] It can be seen from the above embodiments that the disease detection model is used to automatically detect the pathology report to assist doctors in making more accurate diagnosis results based on the pathology report.
[0076] Based on the inventive concept of the above embodiments, the present application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the above embodiments when executing the computer program. Figure 7 Provide detailed explanation.
[0077] like Figure 7 As shown, it shows the electronic device of the present application, which may specifically include a processor 110 and a memory 120. The memory 120 is coupled to the processor 110.
[0078] The processor 110 is used to control the operation of the electronic device. The processor 110 may also be referred to as a CPU (Central Processing Unit). The processor 110 may be an integrated circuit chip having the ability to process signals. The processor 110 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. A general-purpose processor may be a microprocessor or the processor 110 may also be any conventional processor, etc.
[0079] The memory 120 is used to store computer programs, and may be a RAM, a ROM, or other types of storage terminals. Specifically, the memory 120 may include one or more computer-readable storage media, which may be non-transitory or transient. The memory 120 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage terminals, flash memory storage terminals. In some embodiments, the non-transitory computer-readable storage medium in the memory 120 is used to store at least one program code.
[0080] The processor 110 is used to execute the computer program stored in the memory 120 to implement the methods described in the various method embodiments of the present application.
[0081] In some embodiments, the electronic device may further include: a peripheral terminal interface 130 and at least one peripheral terminal. The processor 110, the memory 120 and the peripheral terminal interface 130 may be connected via a bus or a signal line. Each peripheral terminal may be connected to the peripheral terminal interface 130 via a bus, a signal line or a circuit board. Specifically, the peripheral terminal includes: at least one of a radio frequency circuit 140, a display screen 150, an audio circuit 160 and a power supply 170.
[0082] The peripheral terminal interface 130 may be used to connect at least one peripheral terminal related to I / O (Input / output) to the processor 110 and the memory 120. In some embodiments, the processor 110, the memory 120, and the peripheral terminal interface 130 are integrated on the same chip or circuit board; in some other implementations, any one or two of the processor 110, the memory 120, and the peripheral terminal interface 130 may be implemented on a separate chip or circuit board, which is not limited in this embodiment.
[0083] The radio frequency circuit 140 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The radio frequency circuit 140 communicates with the communication network and other IoT devices through electromagnetic signals, and the radio frequency circuit 140 is the communication circuit of the electronic device. The radio frequency circuit 140 converts the electrical signal into an electromagnetic signal for transmission, or converts the received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 140 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, an operator identity module card, and the like. The radio frequency circuit 140 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes but is not limited to: the World Wide Web, a metropolitan area network, an intranet, various generations of mobile communication networks (2G, 3G, 4G and 5G), a wireless local area network and / or a WiFi (Wireless Fidelity) network. In some embodiments, the radio frequency circuit 140 may also include circuits related to NFC (Near Field Communication), which is not limited in this application.
[0084] The display screen 150 is used to display the UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When the display screen 150 is a touch display screen, the display screen 150 also has the ability to collect touch signals on the surface or above the surface of the display screen 150. The touch signal can be input to the processor 110 as a control signal for processing. At this time, the display screen 150 can also be used to provide virtual buttons and / or virtual keyboards, also known as soft buttons and / or soft keyboards. In some embodiments, the display screen 150 can be one, set on the front panel of the electronic device; in other embodiments, the display screen 150 can be at least two, respectively set on different surfaces of the electronic device or in a folding design; in other embodiments, the display screen 150 can be a flexible display screen, set on a curved surface or a folding surface of the electronic device. Even, the display screen 150 can also be set to a non-rectangular irregular shape, that is, a special-shaped screen. The display screen 150 can be made of materials such as LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode, organic light-emitting diode).
[0085] The audio circuit 160 may include a microphone and a speaker. The microphone is used to collect sound waves from the operator and the environment, and convert the sound waves into electrical signals and input them into the processor 110 for processing, or input them into the radio frequency circuit 140 to achieve voice communication. For the purpose of stereo acquisition or noise reduction, there can be multiple microphones, which are respectively arranged in different parts of the electronic device. The microphone can also be an array microphone or an omnidirectional acquisition microphone. The speaker is used to convert the electrical signal from the processor 110 or the radio frequency circuit 140 into sound waves. The speaker can be a traditional film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can not only convert the electrical signal into sound waves audible to humans, but also convert the electrical signal into sound waves inaudible to humans for purposes such as ranging. In some embodiments, the audio circuit 160 may also include a headphone jack.
[0086] The power supply 170 is used to power various components in the electronic device. The power supply 170 can be an alternating current, a direct current, a disposable battery, or a rechargeable battery. When the power supply 170 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery that is charged through a wired line, and a wireless rechargeable battery is a battery that is charged through a wireless coil. The rechargeable battery can also be used to support fast charging technology.
[0087] For a detailed description of the functions and execution processes of each functional module or component in the electronic device embodiment of the present application, reference can be made to the description in the above-mentioned method embodiments of the present application, which will not be repeated here.
[0088] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the various embodiments of the electronic devices described above are only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.
[0089] 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, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0090] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0091] Based on the inventive concept of the above embodiment, the present application also provides a computer-readable storage medium storing a computer program, which can be executed by a processor to implement the above projection area calculation method. Figure 8 The execution process of the above-mentioned embodiment in a computer-readable storage medium is described.
[0092] like Figure 8 As shown, it shows the computer-readable storage medium of the present application. If the above-mentioned 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 200. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions / computer programs to enable an IoT device (which can be a personal computer, server, or network terminal, etc.) or a processor (processor) to perform all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: various media such as a USB flash drive, 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, and electronic terminals such as computers, mobile phones, laptops, tablet computers, cameras, etc. having the above-mentioned storage medium.
[0093] The description of the execution process of the program data in the computer-readable storage medium can refer to the description in the above-mentioned method embodiments of the present application, which will not be repeated here.
[0094] The above descriptions are merely 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.
[0095] Those skilled in the art will appreciate that, in the above method of specific implementation, the order in which the steps are written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of the steps should be determined by their functions and possible internal logic.
Claims
1. A pathology report annotation method, characterized in that: The method comprises: Obtain multiple pathology reports corresponding to the target disease in a first preset year; Extracting a plurality of high-frequency words from a plurality of the pathology reports corresponding to the target disease according to a preset standard, and forming a high-frequency word list with the plurality of the high-frequency words; Acquire multiple keywords in the high-frequency word list and form a keyword list, wherein the keywords are obtained by manually screening multiple high-frequency words in the high-frequency word list; Performing multiple iterative screening on all the keywords in the keyword table, and constructing a target keyword table corresponding to the target disease based on the results of the multiple iterative screening; Based on the target keyword table corresponding to the target disease, the acquired pathology report corresponding to the target disease to be labeled is labeled.
2. The method according to claim 1, characterized in that The target keyword table corresponding to the target disease is constructed in the following manner: Verifying all the keywords in the keyword table word by word using a plurality of pathology reports corresponding to the target disease in the first preset year, and constructing a verification word table corresponding to the target disease based on the verified results; All the keywords in the verification vocabulary are tested word by word using a plurality of pathology reports corresponding to the target disease in a second preset year, and a target keyword table corresponding to the target disease is constructed based on the test results.
3. The method according to claim 2, characterized in that The verification vocabulary corresponding to the target disease is constructed in the following manner: All keywords in the keyword table are verified word by word using multiple pathology reports corresponding to the target disease in the first preset year, and the verified results are manually streamlined, and a verification word table corresponding to the target disease is constructed based on the results after manual streamlining.
4. The method according to claim 1, characterized in that: The acquired pathology report corresponding to the target disease to be labeled is labeled in the following manner: All keywords in the target keyword table corresponding to the target disease are graded according to a preset grade allocation rule, and based on the target keyword table corresponding to the target disease after the grade setting, the acquired pathology report corresponding to the target disease to be marked is marked.
5. The method according to claim 1, characterized in that The target disease is breast nodules.
6. A pathology report annotation system based on the method according to any one of claims 1 to 5, characterized in that: The system comprises: A data acquisition module, used to acquire multiple pathology reports corresponding to the target disease in a first preset year; A high-frequency word extraction module, used to extract a plurality of high-frequency words from a plurality of the pathology reports corresponding to the target disease according to a preset standard, and form a high-frequency word list with the plurality of the high-frequency words; A keyword extraction module, used to obtain multiple keywords in the high-frequency word list and form a keyword list, wherein the keywords are obtained by manually screening multiple high-frequency words in the high-frequency word list; A multiple quality control module, used for performing multiple iterative screening on all the keywords in the keyword table, and constructing a target keyword table corresponding to the target disease based on the results of multiple iterative screening; The marking module is used to mark the acquired pathology report corresponding to the target disease to be marked based on the target keyword table corresponding to the target disease.
7. A method for training a disease detection model, characterized in that: The method comprises: Obtain multiple pathology reports corresponding to breast nodules; Using any one of the methods of claims 1 to 5 to annotate the plurality of pathology reports of the breast nodules; A training set is constructed based on the multiple pathological reports of the labeled breast nodules, and the breast nodule detection model is iteratively trained multiple times using the training set until convergence.
8. An application method of a disease detection model, characterized in that: The method comprises: Obtained pathology reports for testing; The disease detection model trained by the method as described in claim 7 is used to obtain the detection result based on the pathology report to be detected.
9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 5, 7 and 8 when executing the computer program.
10. A computer-readable storage medium having program instructions stored thereon, characterized in that: When the program instructions are executed by a processor, the method described in any one of claims 1-5, 7, and 8 is implemented.