Physical examination report structured analysis method and device, electronic equipment and storage medium

Abstract

The invention provides a structural analysis method and device for a physical examination report, electronic equipment and a storage medium. The method comprises the steps of determining a textbox position and text content in the physical examination report; determining target classification information corresponding to the key information in the physical examination report according to the textbox position and the text content; and generating a structured analysis result of the physical examination report according to the textbox position, the text content and the target classification information. According to the method and the device, the structured analysis mode is more flexible, the physical examination reports of complex and changeable styles can be effectively covered, and the structured analysis effect of the physical examination reports is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of medical health technology, and in particular to a method, device, electronic device and storage medium for structured analysis of a physical examination report. Background Art

[0002] A physical examination report is a document in a certain format that is generated based on the data of the examinee's physical response after a physical examination of the examinee through medical means and methods. Through the physical examination report, people can have a preliminary understanding of their health status and detect diseases and health risks at an early stage. The structured analysis of the physical examination report can help medical service personnel collect health information, establish health records, evaluate and predict health trends, specify and implement health plans, and track health management, etc.

[0003] In the related art, the physical examination report is usually structured and parsed based on optical character recognition (OCR) detection and simple rules.

[0004] In this way, the structured analysis method is not flexible enough and cannot effectively cover complex and varied physical examination reports, resulting in poor structured analysis of physical examination reports. Summary of the invention

[0005] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0006] To this end, the purpose of the present disclosure is to propose a method, device, electronic device, non-transitory computer-readable storage medium storing computer instructions, and computer program product for structured parsing of physical examination reports, so that the structured parsing method is more flexible, can effectively cover complex and varied styles of physical examination reports, and improve the structured parsing effect of physical examination reports.

[0007] To achieve the above-mentioned objectives, the first aspect of the present disclosure proposes a method for structured parsing of a physical examination report, comprising: determining a text box position and text content in the physical examination report; determining target classification information corresponding to key information in the physical examination report based on the text box position and text content; and generating a structured parsing result of the physical examination report based on the text box position and text content, and the target classification information.

[0008] To achieve the above-mentioned purpose, a physical examination report structured parsing device proposed in an embodiment of the second aspect of the present disclosure includes: a first determination module, used to determine the text box position and text content in the physical examination report; a second determination module, used to determine the target classification information corresponding to the key information in the physical examination report according to the text box position and text content; and a generation module, used to generate a structured parsing result of the physical examination report according to the text box position and text content, and the target classification information.

[0009] The electronic device proposed in the third aspect of the present disclosure includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the structured parsing method for the physical examination report proposed in the first aspect of the present disclosure is implemented.

[0010] The fourth aspect embodiment of the present disclosure proposes a non-temporary computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the structured parsing method of the physical examination report proposed in the first aspect embodiment of the present disclosure.

[0011] The fifth aspect of the present disclosure provides a computer program product. When the instructions in the computer program product are executed by a processor, the method for structured parsing of a physical examination report provided in the first aspect of the present disclosure is performed.

[0012] The physical examination report structured parsing method, device, electronic device, non-transient computer-readable storage medium storing computer instructions, and computer program product provided by the present disclosure first determine the text box position and text content in the physical examination report, and then determine the target classification information corresponding to the key information in the physical examination report based on the text box position and text content. By incorporating the relationship between different text box positions and text contents into the recognition process of structured categories, it can have higher robustness and generalization ability. For physical examination reports of various types and complex scenarios, more accurate physical examination report structured parsing results can be obtained, thereby making the structured parsing method more flexible, able to effectively cover complex and varied styles of physical examination reports, and improving the structured parsing effect of physical examination reports.

[0013] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description or learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0015] Figure 1 It is a flowchart of a method for structured analysis of a physical examination report proposed in an embodiment of the present disclosure;

[0016] Figure 2 is a flow chart of a method for structured analysis of a physical examination report proposed in another embodiment of the present disclosure;

[0017] Figure 3 is a flow chart of a method for structured analysis of a physical examination report proposed in another embodiment of the present disclosure;

[0018] Figure 4 is a schematic diagram of correction of initial classification information in an embodiment of the present disclosure;

[0019] Figure 5 is a schematic diagram of a structured analysis process in an embodiment of the present disclosure;

[0020] Figure 6 is a schematic diagram of a structured analysis result in an embodiment of the present disclosure;

[0021] Figure 7 It is a structural schematic diagram of a physical examination report structured analysis device proposed in an embodiment of the present disclosure;

[0022] Figure 8 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0023] Embodiments of the present disclosure are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present disclosure, and are not to be construed as limitations of the present disclosure. On the contrary, the embodiments of the present disclosure include all changes, modifications, and equivalents that fall within the spirit and connotation of the appended claims.

[0024] Figure 1 It is a flowchart of a method for structured analysis of a physical examination report proposed in an embodiment of the present disclosure.

[0025] This embodiment takes the physical examination report structured parsing method as an example configured in a physical examination report structured parsing device. In this embodiment, the physical examination report structured parsing method can be configured in a physical examination report structured parsing device, and the physical examination report structured parsing device can be set in a server, or can also be set in an electronic device, without limitation. This embodiment takes the physical examination report structured parsing method as an example configured in an electronic device. Among them, electronic devices such as smart phones, tablet computers, personal digital assistants, e-books and other hardware devices with various operating systems.

[0026] It should be noted that the execution subject of the embodiment of the present disclosure may be, for example, a central processing unit (CPU) in a server or electronic device in terms of hardware, or may be, for example, a related background service in a server or electronic device in terms of software, without limitation.

[0027] like Figure 1 As shown, the structured parsing method of the physical examination report includes:

[0028] S101: Determine the text box position and text content in the physical examination report.

[0029] Among them, the physical examination report may include the basic information of the examinee, information on physical examination items and the overall examination conclusion. With the improvement of social living standards, people are paying more and more attention to the physical health of themselves and their families, and the stock of physical examination report data is also increasing day by day. As a result, the layout style of the basic information of the examinee, information on physical examination items and the overall examination conclusion in the physical examination report is becoming increasingly complex and changeable.

[0030] Among them, the basic information of the examinee, physical examination item information and general examination conclusion in the physical examination report may exist in the form of pictures, texts, characters, etc. In the embodiment of the present disclosure, the text box in the physical examination report can be detected. The text box can be understood as a detection box containing a paragraph of text, for example, a rectangular detection box containing a paragraph of text. The position of the text box can be used to describe the position of the text box in the physical examination report, and the text content can refer to the content of the text framed by the text box, such as the text morphological content, the semantic content of the text, etc., and there is no limitation on this.

[0031] In the disclosed embodiment, the physical examination report can be detected and identified to identify a text box containing text, and then the position of the text box in the physical examination report is determined as the text box position, and the specific semantic content of the text is determined as the text content, without limitation.

[0032] Optionally, in some embodiments, when executing the step of determining the position of a text box in a physical examination report, a report image corresponding to the physical examination report may be generated, and at least one text box in the report image may be identified, as well as the initial position of the text box and the relative position between different text boxes. The initial position and relative position of the text box are collectively used as the text box position. The resulting text box position can not only indicate the position of the text box itself, but can also be used to indicate the relative position between different text boxes, which can be beneficial to layout recovery during the structured parsing process of the physical examination report and help improve the structured parsing effect of the physical examination report.

[0033] The initial position of the text box indicates the position of the text box itself, such as the position of the text box in the physical examination report; the relative position between different text boxes indicates the relative position between different text boxes.

[0034] In some embodiments, a reference coordinate system may be set for the physical examination report to determine the position coordinates of the text box in the physical examination report in the reference coordinate system. The position coordinates may be an optional example of the initial position of the text box, which is not limited thereto.

[0035] Of course, any other possible manner may be used to identify the position of the text box in the physical examination report, and the position of the text box in the physical examination report may be used as the initial position of the text box, and there is no limitation to this.

[0036] In some embodiments, the physical examination report may contain at least one page, and each page of the physical examination report can be imaged to generate a report image corresponding to each page of the physical examination report. For example, the physical examination report may be in a document form, and the physical examination report in the document form can be processed into an image, which can be called a report image, and then the text box of the report image can be detected and identified, without limitation.

[0037] In some embodiments, the detection and identification of text boxes in a report image can be performed by identifying a detection box (text box) containing text from the report image based on OCR recognition rules, and then determining the overall position of the report image in a reference coordinate system, and analyzing the local position of the text box in the reference coordinate system, and determining the initial position of the text box and the relative position between different text boxes based on the local position and the overall position, without limitation.

[0038] In other embodiments, the report image can also be input into a text box detection model, and the report image can be detected and identified based on the text box detection model to determine the initial position of the text box in the report image and the relative position between different text boxes. There is no limitation on this.

[0039] In other embodiments, the input physical examination report can be converted into image format page by page, and then each page can be separately input into an open source OCR detection and recognition model for OCR detection and recognition. The OCR detection and recognition model can, for example, be implemented based on PaddleOCR, which is an open source OCR tool library based on Paddle. Paddle is an open source deep learning platform. Then, based on the open source OCR detection and recognition model, OCR detection and recognition is performed on the physical examination report in image format to determine the initial position of the text box in the report image, and the relative position between different text boxes, without any limitation.

[0040] In some embodiments of the present disclosure, in the process of determining the text content in a physical examination report, the text content of a text box in a report image can be identified based on the initial position and relative position of the text box, so that the text content can be quickly located and the accuracy of text content detection and recognition can be improved.

[0041] In other embodiments of the present disclosure, the relative positions of different text boxes are determined. The different text boxes may be different text boxes that have an associated relationship. For example, a page of a physical examination report may contain multiple text boxes, and the text contents contained in some text boxes may be related. In this case, the some text boxes may be regarded as having an associated relationship, while the remaining text boxes containing irrelevant text contents may not have an associated relationship. In this case, when the text content in the text box is identified based on the relative positions of different text boxes and the positions of the text boxes, it can be ensured that the identified text content has a high reference value, thereby improving the detection effect of the text content.

[0042] S102: Determine target classification information corresponding to key information in the physical examination report according to the text box position and text content.

[0043] In determining the text box position and text content in the physical examination report, target classification information corresponding to key information in the physical examination report can be determined based on the text box position and text content.

[0044] Among them, key information refers to the relatively critical information in the physical examination report, such as age, gender, contact information, physical examination items and other information.

[0045] Among them, the target classification information can be extracted from the key information analysis based on the classification category of the key information. The target classification information refers to part of the key information related to the classification category. For example, when performing structured analysis on the physical examination report, the key information in the physical examination report is structured and classified to obtain the classification categories: physical examination item name (table_entryName_VALUE), physical examination item result (table_result_VALUE), physical examination item reference range (table_reference_val_VALUE), physical examination item unit (table_unit_VALUE), physical examination item abbreviation (table_abbreviation_VALUE), methodology (table_methodology_VALUE), etc. Then, part of the information related to the classification category is extracted from the text content and key information analysis as the target classification information, and there is no restriction on this.

[0046] In some embodiments, the text box position, text content, and report image may be input into a classification information recognition model to obtain target classification information corresponding to key information in the physical examination report identified by the classification information recognition model, without limitation.

[0047] In some other embodiments, the text box position may be first referred to to perform semantic and attribute detection on the text content to determine the semantics and attributes of the text content, and then, the key information may be identified in combination with the semantics and attributes of the text content, and the key information may be classified according to a preset classification method to obtain target classification information of the key information, without limitation.

[0048] S103: Generate a structured parsing result of the physical examination report according to the text box position and text content, as well as the target classification information.

[0049] Among them, structured analysis results refer to the analysis results that can present the physical examination report in an intuitive and hierarchical form.

[0050] After determining the target classification information corresponding to the key information in the physical examination report according to the text box position and the text content, a structured parsing result of the physical examination report is generated according to the text box position and the text content, as well as the target classification information.

[0051] In some embodiments, the layout recovery in the structured parsing process of the physical examination report can be achieved by referring to the text box position to obtain the layout frame, and the text content and target classification information can be filled into the layout frame to form an intuitive result with a certain hierarchical structure as a structured parsing result, and there is no limitation on this.

[0052] In other embodiments, the text box position and text content, as well as the target classification information, may be processed for structured output to output structured parsing results, without limitation.

[0053] In some other embodiments, the physical examination report can also be disassembled according to the corresponding reference range and other information, and the disassembled content can be divided according to the connector, split into separate upper and lower limits and filled into the corresponding structured fields. At the same time, the abnormal conditions of the indicators are calculated, and the indicators greater than the upper limit or lower than the lower limit are marked in the abnormal field. The content of the conclusion is appropriately spliced ​​according to the position of the text box to enhance the readability of the structured output results, and there is no restriction on this.

[0054] Therefore, in this embodiment, since the text box position and text content in the physical examination report are first determined, and then the target classification information corresponding to the key information in the physical examination report is determined based on the text box position and text content, by incorporating the relationship between different text box positions and text content into the recognition process of structured categories, it can have higher robustness and generalization ability. For physical examination reports of various types and complex scenarios, more accurate physical examination report structured analysis results can be obtained. Therefore, in the disclosed embodiment, the structured analysis method is made more flexible, which can effectively cover complex and varied styles of physical examination reports and improve the structured analysis effect of physical examination reports.

[0055] Figure 2 It is a flowchart of a method for structured analysis of a physical examination report proposed in another embodiment of the present disclosure.

[0056] like Figure 2 As shown, the structured parsing method of the physical examination report includes:

[0057] S201: Determine a text box position and text content in a physical examination report, wherein the number of lines of the text content is at least one line.

[0058] That is to say, the number of lines of the text content in the physical examination report can be one line or multiple lines.

[0059] In some embodiments, different lines of text content may be recognized separately, or several lines of text content may be recognized in parallel, without limitation.

[0060] S202: Identify key information contained in each line of text content according to the position of the text box.

[0061] For example, if the text content consists of only one line, then based on the position of the text box, the line of text content can be located in the physical examination report to identify the key information contained in the line of text content. If the text content consists of multiple lines, then based on the position of the text box, several lines of text content can be located in the physical examination report to identify the key information contained in several lines of text content. There is no limitation to this.

[0062] In some embodiments of the present disclosure, when executing the step of identifying the key information contained in each line of text content according to the text box position, the text box position and each line of text content can be input into a key information extraction model to obtain the key information output by the key information extraction model, wherein the key information extraction model has learned the mapping relationship between the text box position, each line of text content, and the key information, which can effectively improve the accuracy and efficiency of key information recognition.

[0063] Among them, the key information extraction model can be a neural network model, machine learning model, etc. in artificial intelligence, and there is no restriction on this.

[0064] In some embodiments, a neural network model or a machine learning model can be pre-trained based on the text box position of the sample, the text content of each line of the sample, and the key information of the sample. When the trained neural network model or machine learning model converges, the trained neural network model or machine learning model can be used as a key information extraction model to ensure that the key information extraction model can learn the mapping relationship between the text box position, the text content of each line of the sample, and the key information.

[0065] For example, the input physical examination report can be converted into image format page by page, and then each page can be input into the key information extraction model separately. The key information extraction model can be implemented based on PaddleOCR to obtain the key information output by the key information extraction model. There is no restriction on this.

[0066] S203: Determine target classification information according to the content of each line of text and corresponding key information.

[0067] In determining the text box position and text content in the physical examination report, target classification information corresponding to key information in the physical examination report can be determined based on the text box position and text content.

[0068] Among them, key information refers to the relatively critical information in the physical examination report, such as age, gender, contact information, physical examination items and other information.

[0069] Among them, the target classification information can be extracted from the key information analysis based on the classification category of the key information. The target classification information refers to part of the key information related to the classification category. For example, when performing structured analysis on the physical examination report, the key information in the physical examination report is structured and classified to obtain the classification categories: physical examination item name (table_entryName_VALUE), physical examination item result (table_result_VALUE), physical examination item reference range (table_reference_val_VALUE), physical examination item unit (table_unit_VALUE), physical examination item abbreviation (table_abbreviation_VALUE), methodology (table_methodology_VALUE), etc. Then, part of the information related to the classification category is extracted from the text content and key information analysis as the target classification information, and there is no restriction on this.

[0070] In some embodiments of the present disclosure, when executing the step of determining target classification information based on each line of text content and corresponding key information, at least one standardized classification category can be obtained, and the classification category corresponding to each line of text content and the corresponding key information can be determined, and each line of text content and the corresponding key information can be classified according to the corresponding classification category to obtain the target classification information, which can achieve unified and standardized structuring and effectively avoid ambiguous problems that may be encountered during the structured parsing of the physical examination report, thereby further improving the accuracy of the structured parsing of the physical examination report.

[0071] Among them, at least one classification category that has been standardized can be exemplified as follows:

[0072] 1. Set other to mark unnecessary information in the physical examination report, such as page numbers, directories, etc.

[0073] 2. For attributes that appear in pairs, set the _QUESTION and _ANSWER attributes to pair them.

[0074] 3. For the properties of the table, set the header attributes _HEAD and _VALUE, the table title _TITLE, the general inspection conclusion _ALL, the sub-inspection conclusion detail prefix, etc.

[0075] Based on the massive data sets of multiple physical examination institutions, multiple classification categories can be divided, as listed below:

[0076] Other;

[0077] table_reference_val_VALUE;

[0078] detail_conclusion_CONCLUSION;

[0079] detail_conclusion_medical_interpretation_HEAD;

[0080] table_methodology_VALUE;

[0081] reportCode_QUESTION;

[0082] table_entryName_HEAD;

[0083] table_notes_VALUE;

[0084] reportCode_ANSWER;

[0085] table_abbreviation_VALUE;

[0086] total_conclusion_VALUE;

[0087] table_methodology_HEAD;

[0088] table_TITLE.

[0089] After determining the multiple standardized classification categories as described above, the classification category corresponding to each line of text content and the corresponding key information can be determined, and then, partial information related to the classification category is extracted from the text content and key information analysis as target classification information, without limitation.

[0090] For example, if the number of lines of text content is multiple lines, key information extraction and classification of key information can be performed line by line. The text content in the original physical examination report can be extracted line by line according to the position of the text box, and the text content of each line is classified into categories in combination with the corresponding key information. Set the problem identification, header identification, conclusion identification and other information in the classification category as structured flag conditions to classify the text content and the corresponding key information into the corresponding classification category. Among them, the text content corresponding to the age_ANSWER key information, the structured output is the age in the physical examination report (an optional example of a target classification information), the text content corresponding to the name_ANSWER key information, the structured output is the name in the physical examination report (an optional example of a target classification information), the text content corresponding to the phone_ANSWER key information, the structured output is the phone number in the physical examination report (an optional example of a target classification information), and the other *__ANSWER fields are similar. For specific physical examination items, set table_TITLE as the table start condition, and then extract rows according to the physical examination item name (table_entryName_VALUE), physical examination item result (table_result_VALUE), physical examination item reference range (table_reference_val_VALUE), physical examination item unit (table_unit_VALUE), physical examination item abbreviation (table_abbreviation_VALUE), methodology (table_methodology_VALUE), etc. to identify and classify key information, and then generate structured results.

[0091] S204: Generate a structured parsing result of the physical examination report according to the text box position and text content, as well as the target classification information.

[0092] In this embodiment, since the text box position and text content in the physical examination report are first determined, and then the target classification information corresponding to the key information in the physical examination report is determined according to the text box position and text content, by incorporating the relationship between different text box positions and text content into the recognition process of structured categories, it can have higher robustness and generalization ability. For physical examination reports with a wide variety of types and complex scenes, a more accurate structured parsing result of the physical examination report can be obtained. It also makes the structured parsing method more flexible, can effectively cover physical examination reports of complex and changeable styles, and improve the structured parsing effect of the physical examination report. When executing the step of determining the target classification information according to each line of text content and corresponding key information, it can be to obtain at least one standardized classification category, and determine the classification category corresponding to each line of text content and corresponding key information, and classify each line of text content and corresponding key information according to the corresponding classification category to obtain the target classification information, which can achieve unified and standardized structuring, effectively avoid the ambiguous problems that may be encountered in the process of structured parsing of the physical examination report, thereby further improving the accuracy of the structured parsing of the physical examination report.

[0093] Figure 3 It is a flowchart of a method for structured analysis of a physical examination report proposed in another embodiment of the present disclosure.

[0094] like Figure 3 As shown, the structured parsing method of the physical examination report includes:

[0095] S301: Determine the text box position and text content in the physical examination report.

[0096] S302: Determine initial classification information corresponding to key information in the physical examination report according to the text box position and text content.

[0097] S303: Standardize the initial classification information to obtain target classification information.

[0098] In this embodiment, it is also possible to support standardization of the classification results in the structured parsing process, so that the standardized target classification information can accurately represent the classification of key information and / or text content in the physical examination report.

[0099] In some embodiments of the present disclosure, when performing the step of standardizing the initial classification information to obtain the target classification information, the standard classification information can be obtained, and the similarity value between the initial classification information and the standard classification information can be determined. If the similarity value is greater than the similarity threshold, the standard classification information is used as the target classification information; if the similarity value is less than or equal to the similarity threshold, the initial classification information is used as the target classification information.

[0100] Among them, the standard classification information can be pre-constructed. For example, the standard classification information can be modeled based on massive test reports from multiple physical examination institutions, and a standard library can be constructed. The standard library can contain standard classification information. The standard classification information can be specific, for example, the name of a standard physical examination item, and there is no limitation to this.

[0101] Among them, the initial classification information can be the classification information obtained by initially classifying the text box position and text content and key information. Since the initial classification information may have a certain standard deviation, the initial classification information can be corrected according to several standard classification information in the standard library.

[0102] Among them, the similarity value is used to describe the similarity between the initial classification information and the standard classification information. If the similarity value is high, it means that the initial classification information and the standard classification information are relatively similar. Therefore, the initial classification information can be corrected to the standard classification information. If the similarity value is low, it means that the initial classification information and the standard classification information are not similar. It is determined that the initial classification information may be a new classification information that does not exist in the standard library. At this time, it is not suitable to correct the initial classification information based on the standard classification information in the standard library. The initial classification information can be directly used as the target classification information without restriction.

[0103] The similarity threshold refers to a threshold value of similarity for determining whether to correct the initial classification information based on the standard classification information.

[0104] like Figure 4 As shown, Figure 4 : is a schematic diagram of the correction of the initial classification information in the embodiment of the present disclosure. Among them, the standard classification information is [eyelid], and the initial classification information is [eye face]. If it is determined that the similarity value between [eyelid] and [eye face] is greater than the similarity threshold, it means that the two are the same classification information, and [eye face] can be updated to [eyelid], and [eyelid] is used as the target classification information; if it is determined that the similarity value between [eyelid] and [eye face] is less than or equal to the similarity threshold, it means that the two are not the same classification information, and [eye face] can be used as the target classification information.

[0105] S304: Generate a structured parsing result of the physical examination report according to the text box position and text content, as well as the target classification information.

[0106] like Figure 5 As shown, Figure 5It is a schematic diagram of the structured parsing process in the embodiment of the present disclosure. A physical examination report can be uploaded, and the text box position and text content can be identified based on OCR detection. Then, key information is extracted line by line according to the text box position, and the extracted key information is classified. The initial classification information of the initial classification is standardized according to the standard library to obtain the target classification information, and a structured parsing result of the physical examination report is generated according to the text box position and text content, as well as the target classification information.

[0107] In some embodiments of the present disclosure, when executing the step of generating a structured parsing result of a physical examination report according to the text box position and text content, and target classification information, a structured field may be generated according to the text box position and text content, and reference range information corresponding to the target classification information may be obtained, wherein the target classification information indicates that the corresponding key information and / or text content is a detection item result, and the text content is disassembled according to the reference range information, wherein the disassembled content includes at least: upper limit detection content and / or lower limit detection content, and the upper limit detection content and / or lower limit detection content are filled into the structured field to obtain a structured parsing result, which can effectively improve the expression accuracy and clarity of the structured parsing result.

[0108] Among them, the structured fields are, for example, "platelet count" and "red blood cell distribution width standard deviation".

[0109] The reference range information refers to the possible distribution range of the test value of the test item results.

[0110] In some embodiments of the present disclosure, when executing the step of generating a structured parsing result of a physical examination report according to the text box position and text content, as well as target classification information, the text content can also be spliced ​​according to the text box position to obtain a structured parsing result, wherein the target classification information indicates that the corresponding key information and / or text content is a conclusive result, which can effectively improve the readability of the structured parsing result.

[0111] like Figure 6 As shown, Figure 6 It is a schematic diagram of the structured analysis result in the embodiment of the present disclosure.

[0112] In this embodiment, since the text box position and text content in the physical examination report are first determined, and then the target classification information corresponding to the key information in the physical examination report is determined based on the text box position and text content, by incorporating the relationship between different text box positions and text content into the recognition process of structured categories, it can have higher robustness and generalization ability. For physical examination reports with a wide variety and complex scenarios, more accurate structured parsing results of physical examination reports can be obtained. It also makes the structured parsing method more flexible, can effectively cover complex and changeable styles of physical examination reports, and improves the structured parsing effect of physical examination reports. It can also support the standardization of the classification results in the structured parsing process, so that the standardized target classification information can accurately represent the classification of key information and / or text content in the physical examination report.

[0113] Figure 7 It is a structural diagram of a physical examination report structured analysis device proposed in one embodiment of the present disclosure.

[0114] like Figure 7 As shown, the physical examination report structured analysis device 70 includes:

[0115] The first determination module 701 is used to determine the text box position and text content in the physical examination report.

[0116] The second determination module 702 is used to determine the target classification information corresponding to the key information in the physical examination report according to the text box position and text content.

[0117] The generating module 703 is used to generate a structured parsing result of the physical examination report according to the text box position and text content, as well as the target classification information.

[0118] In some embodiments of the present disclosure, the first determining module 701 is specifically configured to:

[0119] generating a report image corresponding to the physical examination report;

[0120] identifying at least one text box in the report image;

[0121] The initial position of the text box and the relative position between different text boxes are determined, wherein the initial position and the relative position of the text box are used together as the text box position.

[0122] In some embodiments of the present disclosure, the first determining module 701 is specifically configured to:

[0123] According to the initial position and relative position of the text box, the text content of the text box in the report image is recognized.

[0124] In some embodiments of the present disclosure, the number of lines of the text content is at least one line; wherein the second determination module 702 is specifically configured to:

[0125] According to the position of the text box, identify the key information contained in each line of text;

[0126] Determine the target classification information based on the content of each line of text and the corresponding key information.

[0127] In some embodiments of the present disclosure, the second determining module 702 is further configured to:

[0128] The text box position and each line of text content are input into the key information extraction model to obtain the key information output by the key information extraction model, wherein the key information extraction model has learned the mapping relationship between the text box position, each line of text content, and the key information.

[0129] In some embodiments of the present disclosure, the second determining module 702 is further configured to:

[0130] obtaining at least one classification category that has been standardized;

[0131] Determine the classification category corresponding to each line of text and the corresponding key information;

[0132] Classify each line of text content and corresponding key information according to the corresponding classification category to obtain the target classification information.

[0133] In some embodiments of the present disclosure, the second determining module 702 is specifically configured to:

[0134] Determine the initial classification information corresponding to the key information in the physical examination report based on the text box position and text content;

[0135] The initial classification information is standardized to obtain the target classification information.

[0136] In some embodiments of the present disclosure, the second determining module 702 is further configured to:

[0137] Get standard classification information;

[0138] Determining a similarity value between the initial classification information and the standard classification information;

[0139] If the similarity value is greater than the similarity threshold, the standard classification information is used as the target classification information;

[0140] If the similarity value is less than or equal to the similarity threshold, the initial classification information is used as the target classification information.

[0141] In some embodiments of the present disclosure, the generating module 703 is specifically used to:

[0142] Generate structured fields based on text box position and text content;

[0143] Acquiring reference range information corresponding to target classification information, wherein the target classification information indicates that corresponding key information and / or text content is a detection item result;

[0144] Decomposing the text content according to the reference range information, wherein the decomposition content at least includes: upper limit detection content and / or lower limit detection content;

[0145] The upper limit detection content and / or the lower limit detection content are filled into the structured field to obtain a structured parsing result.

[0146] In some embodiments of the present disclosure, the generating module 703 is specifically used to:

[0147] The text content is spliced ​​according to the position of the text box to obtain a structured parsing result, wherein the target classification information indicates that the corresponding key information and / or text content is a conclusive result.

[0148] It should be noted that the above explanation of the method for structured analysis of physical examination reports is also applicable to the device for structured analysis of physical examination reports of this embodiment, and will not be repeated here.

[0149] Therefore, in this embodiment, since the text box position and text content in the physical examination report are first determined, and then the target classification information corresponding to the key information in the physical examination report is determined based on the text box position and text content, by incorporating the relationship between different text box positions and text content into the recognition process of structured categories, it can have higher robustness and generalization ability. For physical examination reports of various types and complex scenarios, more accurate physical examination report structured analysis results can be obtained. Therefore, in the disclosed embodiment, the structured analysis method is made more flexible, which can effectively cover complex and varied styles of physical examination reports and improve the structured analysis effect of physical examination reports.

[0150] Figure 8 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present disclosure is shown. Figure 8 The electronic device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0151] like Figure 8 As shown, the electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0152] The bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor or a local bus using any of a variety of bus structures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnection (PCI) bus.

[0153] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0154] The memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 8 Not shown, often called a "hard drive").

[0155] although Figure 8 Not shown in the figure, a disk drive for reading and writing a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a compact disc read only memory (hereinafter referred to as: CD-ROM), a digital versatile disc read only memory (hereinafter referred to as: DVD-ROM) or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present disclosure.

[0156] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in the memory 28, such program modules 42 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 42 generally perform the functions and / or methods of the embodiments described in the present disclosure.

[0157] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), may also communicate with one or more devices that enable the human body to interact with the electronic device 12, and / or communicate with any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed through an input / output (I / O) interface 22. In addition, the electronic device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 through a bus 18. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0158] The processing unit 16 executes various functional applications and data processing by running the programs stored in the system memory 28, such as implementing the structured analysis method of the physical examination report mentioned in the above embodiment.

[0159] In order to implement the above embodiments, the present disclosure further proposes a non-temporary computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the structured parsing method of the physical examination report proposed in the above embodiments of the present disclosure is implemented.

[0160] In order to implement the above embodiments, the present disclosure further proposes a computer program product. When the instructions in the computer program product are executed by a processor, the method for structured parsing of a physical examination report proposed in the above embodiments of the present disclosure is executed.

[0161] It should be noted that, in the description of the present disclosure, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present disclosure, unless otherwise specified, the meaning of "plurality" is two or more.

[0162] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present disclosure includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0163] It should be understood that the various parts of the present disclosure can be implemented in hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0164] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0165] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0166] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0167] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0168] Although the embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present disclosure. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present disclosure.

Claims

1. A method for structured analysis of a physical examination report, characterized in that: The method comprises: Determine the text box location and text content in the physical examination report; Determining target classification information corresponding to key information in the physical examination report according to the text box position and text content; and A structured parsing result of the physical examination report is generated according to the text box position and text content, and the target classification information.

2. The method according to claim 1, characterized in that Determine the location of the text box in the physical examination report, including: generating a report image corresponding to the physical examination report; identifying at least one text box in the report image; The initial position of the text box and the relative positions between different text boxes are determined, wherein the initial position of the text box and the relative positions are used together as the text box position.

3. The method according to claim 2, characterized in that Determine the text content of the medical examination report, including: The text content of the text box in the report image is identified according to the initial position of the text box and the relative position.

4. The method according to claim 1, characterized in that The number of lines of the text content is at least one line; wherein, determining the target classification information corresponding to the key information in the physical examination report according to the text box position and the text content includes: According to the position of the text box, identifying key information contained in the text content of each line; The target classification information is determined according to the text content of each line and the corresponding key information.

5. The method according to claim 4, characterized in that The step of identifying key information contained in each line of the text content according to the position of the text box includes: The text box position and the text content of each line are input into the key information extraction model to obtain the key information output by the key information extraction model, wherein the key information extraction model has learned the mapping relationship between the text box position, the text content of each line, and the key information.

6. The method according to claim 4, characterized in that Determining the target classification information according to the text content of each line and the corresponding key information includes: obtaining at least one classification category that has been standardized; Determine the classification category corresponding to the text content and corresponding key information in each line; The text content of each line and the corresponding key information are classified according to the corresponding classification category to obtain the target classification information.

7. The method according to claim 1, characterized in that Determining target classification information corresponding to key information in the physical examination report according to the text box position and text content includes: Determining initial classification information corresponding to key information in the physical examination report according to the text box position and text content; The initial classification information is standardized to obtain the target classification information.

8. The method according to claim 7, characterized in that The standardizing the initial classification information to obtain the target classification information includes: Get standard classification information; Determining a similarity value between the initial classification information and the standard classification information; If the similarity value is greater than a similarity threshold, the standard classification information is used as the target classification information; If the similarity value is less than or equal to the similarity threshold, the initial classification information is used as the target classification information.

9. The method according to any one of claims 1 to 8, characterized in that Generating the structured parsing result of the physical examination report according to the text box position and text content, and the target classification information includes: Generate a structured field according to the text box position and text content; Acquire reference range information corresponding to the target classification information, wherein the target classification information indicates that the corresponding key information and / or the text content is a detection item result; Decomposing the text content according to the reference range information, wherein the decomposing content at least includes: upper limit detection content and / or lower limit detection content; The upper limit detection content and / or the lower limit detection content are filled into the structured field to obtain the structured parsing result.

10. The method according to any one of claims 1 to 8, characterized in that Generating the structured parsing result of the physical examination report according to the text box position and text content, and the target classification information includes: The text content is spliced ​​according to the text box position to obtain the structured parsing result, wherein the target classification information indicates that the corresponding key information and / or the text content is a conclusive result.

11. A structured analysis device for a physical examination report, characterized in that: The device comprises: The first determination module is used to determine the text box position and text content in the physical examination report; A second determination module is used to determine target classification information corresponding to the key information in the physical examination report according to the text box position and text content; and A generating module is used to generate a structured parsing result of the physical examination report according to the text box position and text content, and the target classification information.

12. The device according to claim 11, characterized in that The first determining module is specifically configured to: generating a report image corresponding to the physical examination report; identifying at least one text box in the report image; The initial position of the text box and the relative positions between different text boxes are determined, wherein the initial position of the text box and the relative positions are used together as the text box position.

13. The device according to claim 12, characterized in that The first determining module is specifically configured to: The text content of the text box in the report image is identified according to the initial position of the text box and the relative position.

14. The device according to claim 11, characterized in that The number of lines of the text content is at least one line; wherein the second determination module is specifically used to: According to the position of the text box, identifying key information contained in the text content of each line; The target classification information is determined according to the text content of each line and the corresponding key information.

15. The device according to claim 14, characterized in that The second determining module is further used for: The text box position and the text content of each line are input into the key information extraction model to obtain the key information output by the key information extraction model, wherein the key information extraction model has learned the mapping relationship between the text box position, the text content of each line, and the key information.

16. The device according to claim 14, characterized in that The second determining module is further used to: obtaining at least one classification category that has been standardized; Determine the classification category corresponding to the text content and corresponding key information in each line; The text content of each line and the corresponding key information are classified according to the corresponding classification category to obtain the target classification information.

17. The device according to claim 11, characterized in that The second determining module is specifically used to: Determining initial classification information corresponding to key information in the physical examination report according to the text box position and text content; The initial classification information is standardized to obtain the target classification information.

18. The device according to claim 17, characterized in that The second determining module is further used for: Get standard classification information; Determining a similarity value between the initial classification information and the standard classification information; If the similarity value is greater than a similarity threshold, the standard classification information is used as the target classification information; If the similarity value is less than or equal to the similarity threshold, the initial classification information is used as the target classification information.

19. The device according to any one of claims 11 to 18, characterized in that The generation module is specifically used for: Generate a structured field according to the text box position and text content; Acquire reference range information corresponding to the target classification information, wherein the target classification information indicates that the corresponding key information and / or the text content is a detection item result; Decomposing the text content according to the reference range information, wherein the decomposing content at least includes: upper limit detection content and / or lower limit detection content; The upper limit detection content and / or the lower limit detection content are filled into the structured field to obtain the structured parsing result.

20. The device according to any one of claims 11 to 18, characterized in that The generation module is specifically used for: The text content is spliced ​​according to the text box position to obtain the structured parsing result, wherein the target classification information indicates that the corresponding key information and / or the text content is a conclusive result.

21. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

22. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: in, The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 10.

23. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 10.

Images