Laboratory detection report proofreading method and system
By performing text preprocessing, natural language processing and data judgment on laboratory test reports, combined with semantic analysis methods, the problem of low manual proofreading efficiency is solved, and efficient and accurate proofreading of test reports is achieved.
Patent Information
- Application Number
- CN202411922939.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-25
- Publication Date
- 2025-04-18
AI Technical Summary
In the prior art, the proofreading of laboratory test reports relies on manual operations, resulting in low work efficiency and affecting the timeliness of test reports.
A laboratory test report proofreading method is used, including text preprocessing and natural language processing technology for error detection, data judgment and semantic analysis, and finally the report results are displayed.
It improves the efficiency of proofreading of test reports, reduces the time for manual proofreading, enhances the accuracy and standardization of proofreading, reduces costs, and improves the timeliness of release of test reports.
Smart Images
Figure CN119849485A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automated technology for test report verification, and in particular to a method and system for verifying laboratory test reports. Background Technology
[0002] The state has clear requirements for laboratory test report results, requiring that the test report accurately, clearly, explicitly and objectively reflect the test results, and should include all information that the client agrees to, that is necessary to interpret the results, and that is required by the methods used. This requires that the test report data be accurate and reliable, and that the report be standardized and rigorous.
[0003] Currently, the proofreading of test reports from third-party laboratory testing institutions or companies' own laboratories mainly adopts manual proofreading methods. After the initial draft of the report is edited by the report editor, the report editor submits it to relatively experienced report proofreaders through the laboratory information management system or in the form of Word or PDF files. The report proofreaders proofread the report in the laboratory information management system, in Word or PDF electronic files, or on paper. They then make annotations on the report template, wording and description, questionable or incorrect data, incorrect judgments, inconsistencies in information such as images and text, omissions of required information, etc.
[0004] The main problems in the process of proofreading the above test reports are as follows: the proofreading of documents relies on manual proofreading, which is inefficient and affects the timeliness of the test report release. Summary of the Invention
[0005] This invention provides a laboratory test report proofreading method and system, which solves the technical problem that the proofreading of test reports in the prior art relies on manual operation by proofreaders, improves work efficiency, and does not affect the timeliness of test report release.
[0006] This invention provides a method for calibrating laboratory test reports, comprising:
[0007] Extract the laboratory's test report and perform text preprocessing on the test report; wherein, the text preprocessing includes removing irrelevant formatting, word segmentation, and punctuation marks;
[0008] The NPL technology is used to perform error detection on the preprocessed text of the detection report;
[0009] The detection report after error detection is subjected to data judgment; wherein, the data judgment includes judgment and verification of data units in the table, verification of significant digits in the data in the table, one-way and two-way judgment of data size in the table, verification of data in the text, and verification of consistency between data in the text or table and data in the image;
[0010] The detection report, after judging the data through attribute calculation and attribute grammar, is combined with the context for semantic analysis.
[0011] The detection report after semantic analysis will be displayed as a report result.
[0012] Furthermore, before the step of obtaining the laboratory's test report and performing text preprocessing on the test report, the method further includes:
[0013] Obtain the laboratory's test report and determine whether the test report has a fixed template;
[0014] If the laboratory's test reports have a fixed template, then the content of the test reports is matched and verified to see if it conforms to the predetermined template using a text structure analysis tool;
[0015] If the laboratory's test report does not have a fixed template, select the system-recommended template or customize a fixed template, or do not select any template.
[0016] Furthermore, the step of performing data judgment on the detection report after error detection includes:
[0017] Data unit judgment and verification in the table: A unit lookup table is defined by programming to identify and verify whether the data units are correct; the unit lookup table sets the correspondence between standard units and commonly used units;
[0018] Verification of significant digits in the table: The significant digits of the data are determined by writing rounding rules or algorithms; wherein, the rounding rules are written according to the data representation and measurement precision;
[0019] One-way and two-way judgment of data size in the table: Data size is judged by setting up and writing functions to compare data values. For two-way judgments such as "data 1 ≤ test value ≤ data 2", corresponding logical conditions are added to the function for processing. For professional testing laboratories, the testing methods, testing conditions, and testing judgments are relatively fixed. Data processing is done by editing relatively fixed tables to form a proprietary database or table template library. When the test data is incorrect, it can be automatically judged during data input and distinguished with different colors. The two-way judgment is: first data ≤ test value ≤ second data.
[0020] Data proofreading in text: Using natural language processing techniques combined with deep learning models to identify data errors in text;
[0021] Consistency verification between data in text or tables and data in images: By using image recognition technology to convert the data in the image into an editable text format, the identified data is compared with the data in the report.
[0022] Furthermore, the consistency verification between the data in the text or table and the data in the image is achieved by writing automated scripts or using professional data verification tools, including the structure comparison and data comparison functions provided by NineData.
[0023] Furthermore, in the step of displaying the report results of the semantically analyzed detection report, the report results display can be divided into three types: 1) the parts that need to be confirmed after automatic proofreading of the report are marked with color; 2) an updated report is automatically generated based on the proofread and automatically modified content, and the parts that are different from the original detection report are distinguished by color; 3) proofreading annotations are generated based on the erroneous content and the position of the erroneous content in the document to be proofread.
[0024] Furthermore, after the step of displaying the results of the semantically analyzed detection report, the method further includes:
[0025] The report proofreaders manually confirm each annotation and verify the differences between the regenerated test report and the original test report after automatic system proofreading; the manual confirmation includes modification or ignoring, and the differences between the regenerated test report and the original report are distinguished by different colors.
[0026] The present invention also provides a laboratory test report proofreading system, including a report template module, a text processing module, a standard module, a chart module, a report proofreading module, a manual processing module, and a management module, wherein the text processing module includes a preprocessing unit, a detection unit, a data judgment unit, and a semantic analysis unit;
[0027] The preprocessing unit is used to extract the laboratory's test report and perform text preprocessing on the test report; wherein, the text preprocessing includes removing irrelevant formatting, word segmentation, and punctuation marks;
[0028] The detection unit is used to perform error detection on the preprocessed text of the detection report using NPL technology;
[0029] The data judgment unit is used to perform data judgment on the detection report after error detection; wherein, the data judgment includes data unit judgment and verification in the table, verification of significant digits in the data in the table, one-way and two-way judgment of data size in the table, data verification in the text, and consistency verification between data in the text or table and data in the image;
[0030] The semantic analysis unit is used to perform semantic analysis on the detection report after judging the data through attribute calculation and attribute grammar, combined with the context.
[0031] The report proofreading module is used to display the results of the detection report after semantic analysis.
[0032] Furthermore, the management module includes task management, personnel management, query management, statistical reports, access control, system backup management, and email.
[0033] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.
[0034] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.
[0035] The beneficial effects of this invention are as follows:
[0036] This invention extracts laboratory test reports, performs text preprocessing, and then uses NPL technology to detect errors in the preprocessed test report text. After error detection, the test report undergoes data judgment, and semantic analysis is performed on the data-judged test report using attribute calculation and attribute grammar, combined with the context. Finally, the semantically analyzed test report is displayed as a report result. This method can quickly scan and detect errors, combining automated proofreading tools with human proofreading. Human proofreading provides contextual understanding, captures subtle details, and offers personalized feedback. Therefore, combining automated tools and human proofreading achieves optimal results. It enables testing laboratories, quality inspection agencies, metrology units, and medical examination institutions to automatically proofread reports awaiting review, reducing manual review time and human error, improving the accuracy and compliance of proofreading, enhancing proofreading effectiveness, and reducing costs while increasing efficiency. Attached Figure Description
[0037] Figure 1 This is a schematic diagram of a method flow according to an embodiment of the present invention.
[0038] Figure 2 This is a schematic diagram of the device structure according to an embodiment of the present invention.
[0039] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present invention.
[0040] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0041] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0042] like Figure 1As shown, this invention provides a method for proofreading laboratory test reports, enabling automatic proofreading of test reports. The method includes report template matching, text preprocessing, natural language processing, data judgment and processing, context analysis, and report template matching. The specific steps are as follows:
[0043] S1. Extract the laboratory's test report and perform text preprocessing on the test report; wherein, the text preprocessing includes removing irrelevant formatting, word segmentation, and punctuation marks; for example, full-width / half-width conversion, unification of English capitalization, and unification of simplified and traditional Chinese characters, etc.
[0044] S2. Utilize NLP technology to perform error detection on the preprocessed detection report text; employ NLP techniques, such as machine learning and deep learning, to detect errors in the text, including typos, grammatical errors, and collocation errors. BERT's MLM task or Seq2Seq translation task can be used to identify errors.
[0045] S3. Perform data judgment on the detection report after error detection; wherein, the data judgment includes data unit judgment and verification in the table, verification of significant digits in the data in the table, one-way and two-way judgment of data size in the table, data verification in the text, and consistency verification between data in the text or table and data in the image;
[0046] S4. The detection report, after judging the data through attribute calculation and attribute grammar, is combined with the context for semantic analysis.
[0047] Because the test report contains technical terms and domain knowledge, semantic analysis based on the context is necessary. This can be achieved through attribute calculation and attribute grammar; rules can be defined, and tools can automatically perform attribute calculations.
[0048] S5. Display the results of the detection report after semantic analysis.
[0049] In one embodiment, the report results can be displayed in three ways: 1) The parts of the report that need to be confirmed after automatic proofreading are marked with color; 2) An updated report is automatically generated based on the proofread and automatically modified content, and the parts that are different from the original test report are distinguished by color; 3) Proofreading comments are generated based on the erroneous content and the position of the erroneous content in the document to be proofread.
[0050] In one embodiment, after step S5, the method further includes:
[0051] S6. The report proofreader manually confirms (modifies or ignores) each annotation, and confirms the differences between the newly generated test report after automatic system proofreading (the differences from the original report are distinguished by different colors) and the original test report.
[0052] In one embodiment, each testing laboratory typically has its own fixed-style report template. For reports with a fixed template, text structure analysis tools can be used to match and verify whether the report content conforms to the predetermined template. For reports without a fixed template, a system-recommended template can be selected, or a custom fixed template can be created, or no template selection is required. That is, before step S1, the following steps are also included:
[0053] S001. Obtain the laboratory's test report and determine whether the test report has a fixed template;
[0054] S002. If the laboratory's test report has a fixed template, then the content of the test report is matched and checked to see if it conforms to the predetermined template by using a text structure analysis tool.
[0055] S003. If the laboratory's test report does not have a fixed template, select the system-recommended template or customize a fixed template, or do not select any template.
[0056] In one embodiment, in step S3, the data involves data units, significant digits, one-way data judgments (>, ≥, =, <, ≤), two-way data judgments (e.g., data 1 ≤ test value ≤ data 2), etc. The data is presented in text or data tables, and the consistency between the data in text or tables and the data in images, etc. The accuracy and correctness of the data in the test report can be judged by the following methods, or by using natural language understanding tools such as TexSmart to perform fine-grained entity recognition and deep semantic expression to ensure the accuracy of the data.
[0057] Step S3 specifically includes:
[0058] S301. Data Unit Judgment and Verification in Tables: A unit lookup table is defined through programming to identify and verify whether the data units are correct. The unit lookup table sets the correspondence between standard units and commonly used units. For example, for physical quantities such as length and mass, the correspondence between standard units (such as meters and kilograms) and commonly used units (such as centimeters and grams) can be set, and then an automated script can be used to check whether the data units in the report meet expectations.
[0059] S302. Verification of significant digits in table data: The significant digits of data are determined by writing rounding rules or algorithms; wherein the rounding rules are written according to the data representation and measurement precision; for example, for data expressed in scientific notation, the significant digits depend on the number of significant digits in the coefficients, while the exponent does not affect the significant digits. Corresponding verification rules can be written according to the data representation and measurement precision requirements.
[0060] S303, One-way and two-way judgment of data size in a table:
[0061] 1> To determine the size of data, a function is written to compare data values. For example, a threshold or reference value can be set, and logical judgments are used to determine whether the data is "greater than", "less than", or "equal to" a specific value. For two-way judgments such as "data 1 ≤ test value ≤ data 2", corresponding logical conditions are added to the function.
[0062] 2> For professional testing laboratories, the testing methods, testing conditions, and testing judgments are relatively fixed. Data processing is carried out by editing relatively fixed tables to form a proprietary database or table template library. When the test data is incorrect, the testing personnel can see the automatically judged error when entering the data and distinguish the error with different colors.
[0063] S304. Data verification in text: Using natural language processing technology combined with deep learning models, identify data errors in text; for example, train models to identify patterns such as numbers and units, and verify and correct the data in the text.
[0064] S305. Consistency verification between data in text or tables and data in images: By using image recognition technology (such as OCROPOptical Character Recognition), the data in the image is converted into an editable text format, and the recognized data is compared with the data in the report. In one embodiment, this is achieved by writing an automated script or using professional data verification tools, such as the structure comparison and data comparison functions provided by NineData.
[0065] This embodiment extracts laboratory test reports and performs text preprocessing. Then, NPL technology is used to detect errors in the preprocessed test report text. Data judgment is performed on the error-detected test report, and semantic analysis is conducted on the data-judged test report using attribute calculation and attribute grammar, combined with the context. Finally, the semantically analyzed test report is displayed as a report result. This method can quickly scan and detect errors, enabling the combination of automated proofreading tools and human proofreading. Human proofreading provides contextual understanding, captures subtle details, and offers personalized feedback. Therefore, combining automated tools and human proofreading achieves optimal results.
[0066] In one embodiment, such as Figure 2 As shown, the present invention also provides a laboratory test report verification system, including a report template module, a text processing module, a standard module, a chart module, a report verification module, a manual processing module, and a management module. The text processing module includes a preprocessing unit, a detection unit, a data judgment unit, and a semantic analysis unit.
[0067] The text processing module automatically performs text preprocessing and natural language processing on the detection report. In addition to using the system's built-in report proofreading rules, it also allows for the creation of custom proofreading rules, including data processing rules. Its specific unit functions include:
[0068] The preprocessing unit is used to extract the laboratory's test report and perform text preprocessing on the test report; wherein, the text preprocessing includes removing irrelevant formatting, word segmentation, and punctuation marks;
[0069] The detection unit is used to perform error detection on the preprocessed text of the detection report using NPL technology;
[0070] The data judgment unit is used to perform data judgment on the detection report after error detection; wherein, the data judgment includes data unit judgment and verification in the table, verification of significant digits in the data in the table, one-way and two-way judgment of data size in the table, data verification in the text, and consistency verification between data in the text or table and data in the image;
[0071] The semantic analysis unit is used to perform semantic analysis on the detection report after judging the data through attribute calculation and attribute grammar, combined with the context.
[0072] The report proofreading module is used to display the results of the detection report after semantic analysis.
[0073] Standards Module: Supports networked national standards information platforms or self-built standards modules, self-built professional standards folders, or professional document libraries; supports management of testing items, testing indicators, and standard values in product standards; supports management of method standards; provides custom classification management functions for standards; supports management of electronic standard files; supports multiple correspondences between testing items and testing methods; provides functions for copying, batch adding, and batch editing standards; enables batch import of standards through defined Excel templates; and provides functions for replacing and obsolescence of standards.
[0074] The graph and table module: For relatively fixed tables and charts in the test report, a professional graph and table library is formed, which can automatically import or verify the source of the standard library when referencing or verifying data.
[0075] In one embodiment, the report template module includes:
[0076] The test report acquisition unit is used to acquire the laboratory's test reports and determine whether the test reports have a fixed template.
[0077] The first determination unit is used to match and verify whether the content of the test report conforms to the predetermined template by using a text structure analysis tool when the laboratory test report has a fixed template.
[0078] The second judgment unit is used to select a system-recommended template or a custom fixed template when there is no fixed template for laboratory test reports, or to leave the template unselected.
[0079] Report Template Module: Allows for the definition of various report templates, enabling fully automated report creation (supports PDF, Excel, Word, and other formats); supports automatic and manual numbering of reports, as well as the definition of various numbering rules, and can automatically prevent duplicate or skipped numbers; supports online browsing, online approval, and location-based annotation; allows querying of report creation and approval history; supports electronic signatures, electronic seals, and anti-counterfeiting watermarks; supports batch printing of reports; supports recording and limiting the number of print runs; and manages report retrieval and delivery records.
[0080] In one embodiment, the data determination unit includes:
[0081] The first proofreading subunit is used for judging and verifying the units of data in the table: a unit lookup table is defined by programming to identify and verify whether the units of data are correct; wherein, the unit lookup table sets the correspondence between standard units and commonly used units;
[0082] The second proofreading subunit is used to proofread the significant digits of the data in the table: the significant digits of the data are determined by writing rounding rules or algorithms; wherein, the rounding rules are written according to the data representation and measurement accuracy;
[0083] The one-way and two-way judgment sub-units are used for one-way and two-way judgment of data size in the table: the judgment of data size is carried out by setting and writing functions to compare data values, and the two-way judgment, such as "data 1 ≤ test value ≤ data 2", is handled by adding corresponding logical conditions to the function; for professional testing laboratories, the testing methods, testing conditions and testing judgments are relatively fixed, and data processing is carried out by editing relatively fixed tables to form a proprietary database or table template library. When the test data is incorrect, it can be automatically judged at the time of data input and distinguished with different colors;
[0084] The third proofreading subunit is used for data proofreading in the text: it uses natural language processing technology combined with deep learning models to identify data errors in the text;
[0085] The fourth proofreading subunit is used to verify the consistency between data in text or tables and data in images: by using image recognition technology to convert the data in the image into an editable text format, the identified data is compared with the data in the report.
[0086] In one embodiment, in the fourth proofreading subunit, the consistency verification of the data in the text or table with the data in the image is achieved by writing automated scripts or using professional data verification tools, including the structure comparison and data comparison functions provided by NineData.
[0087] In one embodiment, the report proofreading module includes the import of the report to be proofread and the display of proofreading results. The report result display in the report proofreading module can be divided into three types: 1) The parts of the report that need to be confirmed after automatic proofreading are highlighted in color; 2) An updated report is automatically generated based on the proofread and automatically modified content, and the parts that differ from the original test report are distinguished by color; 3) Proofreading annotations are generated based on the erroneous content and its location in the document to be proofread.
[0088] In one embodiment, the manual processing module combines automated proofreading tools with manual proofreading. The manual processing module includes:
[0089] The report proofreaders manually confirm each annotation and verify the differences between the regenerated test report and the original test report after automatic system proofreading; the manual confirmation includes modification or ignoring, and the differences between the regenerated test report and the original report are distinguished by different colors.
[0090] In one embodiment, the management module includes task management, personnel management, query management, statistical reports, access control, system backup management, and email.
[0091] Task Management: Supports task assignment by department, team, or individual; supports multi-level task assignment or forwarding; allows for automatic task assignment through defined task assignment strategies; supports proactive task reminders.
[0092] Personnel Management: Provides basic personnel information management; provides a visual editing interface for personnel file templates; supports personnel leave management; provides personnel electronic key management.
[0093] Query Management: Allows users to customize query conditions for various data objects, including contracts and reports; provides fuzzy search and exact search methods; and allows query results to be exported in Excel format.
[0094] Statistical Reports: Automatically generate statistical reports using custom report templates; statistical reports can be automatically generated on a periodic basis, such as daily, monthly, and annual reports; supports various chart presentation methods, such as bar charts, line charts, and pie charts; provides export functions in Excel and PDF formats.
[0095] Access control: Supports four levels of access control (account, role, permission, and permission scope); supports visual selection and setting of permissions; supports electronic key login.
[0096] System backup management: Provides automatic scheduled backup functions for database files, test reports, etc.
[0097] Email: Can be integrated with workflow and customer management modules to provide email sending functionality; supports management of email sending templates; supports email attachment functionality.
[0098] Each of the above modules, units, and sub-units is used to perform the respective steps in the above laboratory test report verification method. The specific implementation methods are as described in the above method embodiments, and will not be repeated here.
[0099] like Figure 3 As shown, the present invention also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores all data required for the laboratory test report verification process. The network interface allows communication with external terminals via a network connection. The computer program is executed by the processor to implement the laboratory test report verification method.
[0100] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment on which the present application is applied.
[0101] An embodiment of this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described laboratory test report verification methods.
[0102] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media provided in this application and in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0103] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.
[0104] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.
Claims
1. A laboratory test report proofreading method, characterized in that: include: Extract the laboratory's test report and perform text preprocessing on the test report; wherein the text preprocessing includes removing irrelevant formats, word segmentation, and punctuation marks; Using NPL technology to perform error detection on the pre-processed text of the test report; Performing data judgment on the detection report after error detection; wherein the data judgment includes judgment and proofreading of data units in a table, proofreading of the number of significant digits of data in a table, unidirectional and bidirectional judgment of the size of data in a table, proofreading of data in a text, and proofreading of consistency between data in a text or table and data in a picture; Performing semantic analysis on the detection report after data judgment in combination with the context through attribute calculation and attribute grammar; The detection report after semantic analysis is reported and the results are displayed.
2. The laboratory test report proofreading method according to claim 1, characterized in that: Before the step of obtaining the laboratory test report and performing text preprocessing on the test report, the method further includes: Obtain the laboratory's test report and determine whether the test report has a fixed template; If the laboratory's test report has a fixed template, the text structure analysis tool is used to match and proofread whether the content of the test report conforms to the predetermined template; If the laboratory's test report does not have a fixed template, select the system-recommended template or customize a fixed template, or do not check the template.
3. The laboratory test report proofreading method according to claim 1, characterized in that: The step of performing data judgment on the detection report after the error detection includes: Determine and check the data units in the table: define the unit comparison table by programming to identify and check whether the data units are correct; wherein the unit comparison table sets the correspondence between standard units and commonly used units; Checking the number of significant digits of data in a table: determining the number of significant digits of data by writing rounding rules or algorithms; wherein the rounding rules are written according to the representation form and measurement accuracy of the data; One-way and two-way judgment of data size in the table: For the judgment of data size, the function is set to compare the data value, and for the two-way judgment, the corresponding logical conditions are added to the function for processing; for professional testing laboratories, the test methods, test conditions, and test judgments are relatively fixed. Data processing is performed by editing relatively fixed tables to form a proprietary database or table template library. When the test data is wrong, it can be automatically judged during data input and distinguished with different colors; among them, the two-way judgment is the first data ≤ test value ≤ second data; Data proofreading in text: Using natural language processing technology combined with deep learning models to identify data errors in text; Check the consistency between the data in the text or table and the data in the image: Use image recognition technology to convert the data in the image into an editable text format, and compare the recognized data with the data in the report.
4. The laboratory test report proofreading method according to claim 3, characterized in that: The consistency check between the data in the text or table and the data in the picture is achieved by writing an automated script or using professional data verification tools, including the structure comparison and data comparison functions provided by NineData.
5. The laboratory test report proofreading method according to claim 1, characterized in that: In the step of displaying the report results of the test report after semantic analysis, the report result display can be divided into three types: 1) the content part that needs to be confirmed after automatic proofreading of the report is marked with color; 2) based on the proofread and automatically modified content, an updated report is automatically issued, and the parts that are different from the original test report are distinguished by color; 3) proofreading annotations are generated according to the error content and the location of the error content in the document to be proofread.
6. The laboratory test report proofreading method according to claim 1, characterized in that: After the step of presenting the test report after semantic analysis, the method further includes: The report proofreader manually confirms each marked location, and makes a differential confirmation between the test report regenerated after the system's automatic proofreading and the original test report; wherein, the manual confirmation includes modification or neglect, and the differences between the regenerated test report and the original report are distinguished by different colors.
7. A laboratory test report proofreading system, characterized in that: It includes a report template module, a text processing module, a standard module, a chart module, a report proofreading module, a manual processing module, and a management module, wherein the text processing module includes a preprocessing unit, a detection unit, a data judgment unit, and a semantic analysis unit; The preprocessing unit is used to extract the test report of the laboratory and perform text preprocessing on the test report; wherein the text preprocessing includes removing irrelevant formats, word segmentation, and punctuation marks; The detection unit is used to perform error detection on the pre-processed text of the detection report using NPL technology; The data judgment unit is used to perform data judgment on the detection report after error detection; wherein the data judgment includes judgment and proofreading of data units in the table, proofreading of the number of significant digits of data in the table, unidirectional and bidirectional judgment of the size of data in the table, proofreading of data in the text, and consistency proofreading of data in the text or table and data in the picture; The semantic analysis unit is used to perform semantic analysis on the detection report after data judgment in combination with the context through attribute calculation and attribute grammar; The report proofreading module is used to present the report results of the detection report after semantic analysis.
8. The laboratory test report proofreading system according to claim 7, characterized in that: The management module includes task management, personnel management, query management, statistical reports, authority management, system backup management and email.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.