An artificial intelligence management system and management method capable with generating diagnostic structural data

Abstract

This invention discloses an artificial intelligence management system and method for generating diagnostic structured data. It identifies a text diagnostic report issued by a hospital to a user, containing multiple fixed diagnostic fields for filling in various diagnostic information. It also provides a structured data form required by an administrator. Initially, a user-side processing device captures a surface image of the user's text diagnostic report to generate diagnostic image data. Then, an application interface server performs text recognition on the diagnostic image data and converts it into a set of text string information (JSON). This JSON-based diagnostic information is then filled into the information fields of the structured data form by an administrator-side processing server, thereby generating a table file specific to the user.

Description

[Technical Field]

[0001] This invention relates to an artificial intelligence management system and method for generating diagnostic structure data, and in particular to an application programming interface (API) management system and method for converting text data from paper diagnostic reports into text strings. [Previous Technology]

[0002] In the past, to diagnose an injury, one would go to the hospital for relevant examinations and have a doctor issue a diagnosis certificate. The patient would then take the diagnosis certificate to the insurance company to process insurance claims, or when injured by others, the victim would take the diagnosis certificate to the police station or court to file a lawsuit for injury.

[0003] However, in the past, most diagnostic reports were paper documents. If patients had to carry these documents around, they might lose them by accident and have to get new ones, which was quite inconvenient.

[0004] In addition, even though patients can fill in the relevant forms online, the process is quite cumbersome and inconvenient for them.

[0005] Furthermore, the format of the text documents of each hospital is different, making it difficult to digitize and manage the text documents of each hospital, which causes management inconvenience.

[0006] Therefore, how to provide a way to automatically fill in the text data issued by the hospital into the form file, and how to digitize the text data of different formats, will be the key issues and focus of this case. [Summary of the Invention]

[0007] One objective of this invention is to provide an artificial intelligence management system and method that can automatically fill in text data issued by hospitals into a form file, thereby generating diagnostic structure data.

[0008] Another objective of this invention is to provide an artificial intelligence management system and method for generating diagnostic structure data that can digitize text data of different formats.

[0009] The present invention provides a management method for an artificial intelligence management system that generates diagnostic structured data. This method identifies a text diagnostic report issued by a hospital to a user, which has multiple fixed diagnostic fields for filling in multiple diagnostic information. It also provides a structured data form required by an administrator. The structured data form has multiple information fields. The artificial intelligence management system includes a user-end processing device, an application interface server, and an administrator processing server. The user-end processing device includes an image capturing unit and a user-end transmission unit. The application interface server includes an application interface transmission unit and an application interface processing unit. The application interface processing unit includes a character recognition module and a string conversion module. The management processing server includes a management transmission unit, a management processing unit, and a management storage unit. The management processing unit includes an object conversion module. The user transmission unit and the application interface transmission unit are signal-connected to each other, and the application interface transmission unit and the management transmission unit are signal-connected to each other. The management method includes: a) The image capturing unit of the user-end processing device captures the surface image content of the user's text diagnosis report to generate a diagnostic image data, and transmits it to the application interface server through the user-end transmission unit; b) The application interface transmission unit of the application interface server receives the diagnostic image data, the text recognition module of the application interface processing unit performs text recognition on the diagnostic image data, and the string conversion module generates a set of text string information corresponding to the user based on the text recognition content of the text recognition module; and c) The management transmission unit of the management processing server receives the text string information, and the object conversion module of the management processing unit fills the diagnostic content information corresponding to the fixed diagnostic field into the information field of the structured data form based on the text string information (json) corresponding to the diagnostic field, to generate a table file corresponding to the user, and the management storage unit stores the table file.

[0010] In the management method described above, the structured data electronic form 4 includes multiple different categories of medical codes, and step b) includes the following steps: b1) The application interface processing unit determines whether the user's application content information corresponds to one of the medical codes based on the fixed diagnosis field and / or the diagnosis content; and b2) If yes, a corresponding medical code information is generated, and the string conversion module 222 automatically generates the corresponding text string information specific to the user based on the application content information and the medical code information.

[0011] In the management method described above, the structured data form file includes multiple different categories of medical codes, and step b) includes the following steps: b1') The application interface processing unit determines whether the user's application content information corresponds to a part of the medical code based on the application content information and the medical code information; b2') If yes, it recommends a part of the medical code and sends it back to the user terminal processing device; b3') The user terminal processing device transmits the user's selection to the application interface server through the user terminal transmission unit based on the user's selection of one of the recommended medical codes; and b4') Based on the medical code selected by the user, it generates corresponding medical code information, which is then used by the string conversion module to automatically generate the corresponding text string information specific to the user based on the application content information and the medical code information.

[0012] In the management method described above, the application interface processing unit further includes a smart learning module, the management processing server further includes a management operation interface unit, and the management method further includes step d) after step c), when the administrator examines the structured data table electronic file and finds an error, the management operation interface unit receives a correction content input by the administrator, and transmits the correction content to the application interface server through the management transmission unit, and the smart learning module learns and corrects the identification and analysis of the application interface processing unit according to the correction instruction.

[0013] In the management method described above, the application interface processing unit includes an anomaly judgment module, and at least one of the information fields has a preset range value. Step b) further includes the following steps: b1”) When the generated application content information has a numerical value and the corresponding information field has the preset value, the application interface processing unit determines whether the numerical value is within the preset range value; and b2”) If not, the application content information is marked for the administrator to confirm.

[0014] This invention provides an artificial intelligence management system for generating diagnostic structured data. It is used to identify a text diagnostic report issued by a hospital and provided by a user, containing multiple fixed diagnostic fields for filling in multiple diagnostic information. It also provides a structured data form required by an administrator. The structured data form has multiple information fields. The artificial intelligence management system includes a user-end processing device, an application interface server, and an administrator processing server. The user-end processing device includes an image capturing unit and a user-end transmission unit. The image capturing unit captures the surface image content of the user's text diagnostic report to generate diagnostic image data, and the user-end transmission unit transmits the diagnostic image data. The application interface server includes an application interface transmission unit and an application interface processing unit. The application interface transmission unit is signal-connected to the user-end transmission unit and receives data. The diagnostic image data is described. The application interface processing unit includes a text recognition module and a string conversion module. The text recognition module performs text recognition on the diagnostic image data. The string conversion module generates a set of text string information corresponding to the user based on the text recognition content of the text recognition module, which corresponds to the fixed diagnostic fields and the diagnostic content. The administrator processing server includes a management transmission unit, a management processing unit, and a management storage unit. The management transmission unit is connected to the application interface transmission unit by a signal and receives the text string information. The management processing unit includes an object conversion module. The object conversion module fills the diagnostic content information corresponding to the fixed diagnostic fields in the text string information (JSON) into the information fields of the structured data form to generate a table file corresponding to the user. The management storage unit stores the table file.

[0015] In the artificial intelligence management system described above, the structured data electronic form file includes multiple different categories of medical codes. The application interface processing unit determines whether the user's application content information corresponds to one of the medical codes based on the fixed diagnosis field and / or the diagnosis content. If yes, it generates a corresponding medical code information, which is then used by the string conversion module to automatically generate the corresponding text string information specific to the user based on the application content information and the medical code information.

[0016] In the artificial intelligence management system described above, the structured data electronic form file includes multiple different categories of medical codes. The application interface processing unit determines whether the user's application content information corresponds to a portion of the medical codes based on the application content information and the medical code information. If so, it recommends a portion of the medical codes and sends it back to the user terminal processing device. The user terminal processing device transmits the user's selection to the application interface server through the user terminal transmission unit based on the user's selection of one of the recommended medical codes. This allows the application interface processing unit to generate corresponding medical code information based on the medical code selected by the user, so that the string conversion module can automatically generate the corresponding text string information specific to the user based on the application content information and the medical code information.

[0017] In the artificial intelligence management system described above, the application interface processing unit further includes a smart learning module, and the management processing server further includes a management operation interface unit. When the administrator examines the structured data table electronic file and finds an error, the management operation interface unit receives a correction content input by the administrator and transmits the correction content to the application interface server through the management transmission unit. The smart learning module learns and corrects the identification and analysis of the application interface processing unit according to the correction instruction.

[0018] In the artificial intelligence management system described above, the application interface processing unit includes an anomaly judgment module, and at least one of the information fields has a preset range value. When the generated application content information has a numerical value and the corresponding information field has the preset value, the application interface processing unit determines whether the numerical value is within the preset range value. If not, the application content information is marked for the administrator to confirm.

[0019] The present invention provides an artificial intelligence management system and management method for generating diagnostic structure data. It converts text diagnostic reports into a set of text string information through text recognition. The administrator's processing server only needs to receive the text string information and automatically fill it into the information field to automatically generate a table file. Through the recognized and converted text string information, even text diagnostic reports of different formats issued by different hospitals can automatically generate a unified table file to improve the efficiency of digital management and achieve all the above objectives.

Implementation Method

[0021] A first preferred embodiment of the artificial intelligence management system and management method for generating diagnostic structured data of the present invention is shown in Figures 1 and 2. The artificial intelligence management system is for identifying a text diagnostic report 4 issued by a hospital and provided by a user, which has multiple fixed diagnostic fields and multiple diagnostic content information 41 respectively filled in, and for providing a structured data form 5 required by an administrator. The structured data form 5 has multiple information fields 51. In this example, the administrator may be an insurance company, or a judge or prosecutor of a court, and can perform digital generation, conversion and management of the text diagnostic report 4 through the artificial intelligence management system.

[0022] The artificial intelligence management system includes a user terminal processing device 1, an application interface server 2, and a management processing server 3. The user terminal processing device 1 includes an image capture unit 11 and a user terminal transmission unit 12. The application interface server 2 includes an application interface transmission unit 21 and an application interface processing unit 22. The application interface processing unit 22 includes a character recognition module 221 and a string conversion module 222. The management processing server 3 includes a management transmission unit 31, a management processing unit 32, and a management storage unit 33. The management processing unit 32 includes an object conversion module 321. The user terminal transmission unit 12 and the application interface transmission unit 21 are signal-connected to each other, and the application interface transmission unit 21 and the management transmission unit 31 are signal-connected to each other.

[0023] Referring also to Figure 2, which is a flowchart of this embodiment, initially as in step S01, the image capturing unit 11 of the user terminal processing device 1 captures the surface image content of the user's text diagnostic report 4 to generate diagnostic image data, and transmits it to the application interface server 2 through the user terminal transmission unit 12. Next, as in step S02, the application interface server 2's application interface transmission unit 21 receives the diagnostic image data, and the text recognition module 221 of the application interface processing unit 22 performs text recognition on the diagnostic image data, and then the string conversion module 222 converts the text... The text recognition module 221 recognizes the text content, and generates a set of text string information corresponding to the user for the fixed diagnosis field and diagnosis content. In this example, the text string information is in JSON format. Finally, as in step S03, the management transmission unit 31 of the management processing server 3 receives the text string information, and the object conversion module 321 of the management processing unit 32 fills the diagnosis content information 41 of the fixed diagnosis field corresponding to the text string information (JSON) into the information field 51 of the structured data form 5 to generate a table file corresponding to the user. The management storage unit 32 stores the table file.

[0024] Referring also to Figure 3, since each hospital's text diagnosis report 4 has a different format, the artificial intelligence management system in this case can parse each text diagnosis report 4 according to the diagnosis content information 41 of the corresponding fixed diagnosis field. For example, if the fixed diagnosis field is date of birth, ID card information, address, or diagnosis result, the content filled in the corresponding fixed diagnosis field (including the diagnosis result) can be converted into text string information (JSON) and provided to the management server 3. This allows the management server 3 to directly fill the recognized text string information (JSON) into the information field 51 of the structured data form 5 without considering the format of each text diagnosis report 4, so as to generate the required structured data form 5.

[0025] Referring to Figures 1, 2, 3 and 4, in this embodiment, the structured data form file 4 includes multiple different categories of medical codes. In this example, it may be the International Classification of Diseases (ICD) or the National Health Insurance Surgical Code. The step S02 further includes the following steps: First, as in step S021, the application interface processing unit 22 determines whether the user's application content information corresponds to one of the medical codes based on the fixed diagnosis field and / or diagnosis content. If yes, as in step S22, a corresponding medical code information is generated, which is used by the string conversion module 222 to automatically generate the corresponding user-specific text string information (json) based on the application content information and the medical code information. If no, as in step S23, the application interface processing unit 22 requires the user to fill in the medical code and generates the corresponding medical code information based on the medical code entered by the user, which is used by the string conversion module 222 to automatically generate the corresponding user-specific text string information (json) based on the application content information and the medical code information.

[0026] However, in some embodiments, if multiple medical codes are identified, for example, if the text diagnosis report 4 states "surgery," but the corresponding medical codes include surgeries for different parts of the body, the system can provide multiple medical codes for the user to further select. Referring to Figures 1, 2, 3, and 5, in this embodiment, initially as in step S021', the application interface processing unit 22 determines whether the user's application content information corresponds to a specific medical code based on the application content information and the medical code information. If yes, then proceed to step S022. The system recommends a portion of the medical code and sends it back to the user-side processing device 1. Then, as in step S023, the user-side processing device 1 transmits the user's selection to the application interface server 2 via the user-side transmission unit 12 based on one of the selected medical codes. Then, as in step S024, the system generates corresponding medical code information based on the user's selected medical code, so that the string conversion module 222 can automatically generate a corresponding text string (JSON) for the user based on the application content information and the medical code information. If step S021 determines that the user's application content information does not contain a corresponding portion of the medical code, then as in step S025, the application interface processing unit 22 prompts the user to manually enter the medical code, and generates corresponding medical code information based on the entered medical code, so that the string conversion module 222 can automatically generate a corresponding text string (JSON) for the user based on the application content information and the medical code information.

[0027] A second preferred embodiment of the artificial intelligence management system and management method for generating diagnostic structure data of the present invention is shown in FIG6. In this example, the application interface processing unit 22 further includes an intelligent learning module 223 and an anomaly judgment module 224, and the management processing server 3 further includes a management operation interface unit 34. Referring to FIG7, the step between steps S02 and S03 further includes the following steps, such as step S021”, when the generated application content information has a value and the corresponding information field 51 has a preset value, the application interface processing unit 22 determines whether the value is within the preset range. If not, as in step S022”, the application content information is marked for the administrator to confirm, and step S03 continues. If yes, no marking is performed, and step S03 continues. For example, the preset range for ward fees is NT$2,000 to NT$3,000 per day. However, if the requested amount is NT$4,000, the application interface processing unit 22 will determine that the value exceeds the preset range. The actual situation may be an identification error or an input error by the hospital. Therefore, the application interface processing unit 22 can mark this value so that the administrator can directly confirm whether the marked content is correct when reviewing the form file. The marking method can be color marking or background color marking.

[0028] In addition, after step S03, there is a step S04. When the administrator finds an error while reviewing the electronic structure data spreadsheet, the management interface unit 34 receives a correction input from the administrator and transmits the correction to the application interface server 2 through the management transmission unit 31. The intelligent learning module 223 learns and corrects the identification and analysis of the application interface processing unit 22 based on the correction instructions. For example, when the administrator finds an identification error while searching the electronic structure data spreadsheet, and corrects it through the management interface unit 34, the management interface unit 34 will simultaneously send the correction back to the application interface server 2 for the intelligent learning module 223 to learn, thereby increasing the accuracy of identification and analysis.

[0029] The present invention provides an artificial intelligence management system and management method for generating diagnostic structure data. It uses text recognition to convert text diagnostic reports into a set of text string information. The administrator's processing server only needs to receive the text string information and automatically fill it into the information field to automatically generate a table file. Through the text string information of the recognition and conversion, even if different hospitals issue text diagnostic reports of different formats, a unified table file can be automatically generated.

[0030] Furthermore, the artificial intelligence management system and method of the present invention, which generates diagnostic structure data, can automatically identify the type of medical code in the diagnostic content of the text diagnostic report and automatically generate it in a table file, or provide the user with the most suitable medical code option, reducing the time wasted by the user having to search one by one. In addition, the present invention can also add an anomaly judgment mechanism. When the application interface server finds that the values ​​in the text diagnostic report are abnormal, it can mark the corresponding content, and then the administrator can make a manual judgment, reducing the occurrence of input errors and achieving all the above objectives.

[0031] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make some modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims. [Simplified Explanation of the Diagram]

[0020] Figure 1 is a block diagram of a first preferred embodiment of an artificial intelligence management system and management method for generating diagnostic structure data according to the present invention; Figure 2 is a schematic diagram of converting a text diagnostic report of the artificial intelligence management system of Figure 1 into a structured data form; Figure 3 is a flowchart of the artificial intelligence management method of Figure 1; Figure 4 is a further flowchart of the artificial intelligence management method of Figure 3, illustrating that step S02 further includes steps S021-S023; Figure 5 is a further flowchart of the artificial intelligence management method of Figure 3, illustrating that step S02 further includes steps S021'-S025'; Figure 6 is a block diagram of a second preferred embodiment of an artificial intelligence management system and management method for generating diagnostic structure data according to the present invention; and Figure 7 is a flowchart of the artificial intelligence management method of Figure 6.

Claims

1. A management method for an AI-powered management system that generates structured diagnostic data includes a method for identifying a text diagnostic report issued by a hospital and provided by a user, which has multiple fixed diagnostic fields for filling in multiple diagnostic information. The method also provides a structured data form for an administrator, the structured data form having multiple information fields. The AI-powered management system includes a user-side processing device, an application interface server, and an administrator processing server. The user-side processing device includes an image capturing unit and a user-side transmission unit. The application interface server includes an application interface transmission unit and an application interface processing unit. The application interface processing unit includes a character recognition module and a string conversion module. The management-side processing server includes a management-side transmission unit, a management-side processing unit, and a management-side storage unit. The management-side processing unit includes an object conversion module. The user-side transmission unit and the application interface transmission unit are signal-connected to each other. The application interface transmission unit and the management-side transmission unit are also signal-connected to each other. The management method includes: a) The image capturing unit of the user terminal processing device captures the surface image content of the user's text diagnosis report to generate a diagnostic image data, and transmits it to the application interface server through the user terminal transmission unit; b) The application interface transmission unit of the application interface server receives the diagnostic image data, the text recognition module of the application interface processing unit performs text recognition on the diagnostic image data, and the string conversion module generates a set of text string information corresponding to the user based on the text recognition content of the text recognition module; and c) The management transmission unit of the management terminal processing server receives the text string information, and the object conversion module of the management terminal processing unit fills the information fields of the structured data form according to the diagnostic content information corresponding to the fixed diagnostic fields based on the text string information to generate a table file corresponding to the user, and the management storage unit stores the table file. The structured data form includes multiple different categories of medical codes, and step b) includes the following steps: b1) The application interface processing unit determines whether the user's application content information corresponds to one of the medical codes based on the fixed diagnostic fields and / or the diagnostic content; and b2) If yes, a corresponding medical code information is generated, which is then used by the string conversion module to automatically generate the corresponding string information specific to the user based on the application content information and the medical code information; wherein, the medical codes are International Classification of Diseases or National Health Insurance surgical codes.

2. The management method as described in claim 1, wherein the structured data form includes multiple different categories of medical codes, wherein step b) includes the following steps: b1') the application interface processing unit determines whether the user's application content information corresponds to a portion of the medical codes based on the application content information and the medical code information; b2') if yes, recommends a portion of the medical codes and sends it back to the user terminal processing device; b3') the user terminal processing device transmits the user's selection to the application interface server through the user terminal transmission unit based on the user's selection of one of the recommended medical codes; and b4') based on the medical code selected by the user, generates corresponding medical code information, which is then used by the string conversion module to automatically generate the corresponding string information specific to the user based on the application content information and the medical code information.

3. The management method as described in claim 1, wherein the application interface processing unit further includes a smart learning module, the management processing server further includes a management operation interface unit, and the management method further includes step d) after step c), when the administrator examines the structured data form electronic file and finds an error, the management operation interface unit receives a correction content input by the administrator, and transmits the correction content to the application interface server through the management transmission unit, and the smart learning module learns and corrects the identification and analysis of the application interface processing unit according to the correction instruction.

4. The management method as described in claim 1, wherein the application interface processing unit includes an exception judgment module, and at least one of the information fields has a preset range value, wherein step b) further includes the following steps: b1”) when the generated application content information has a value and the corresponding information field has the preset value, the application interface processing unit determines whether the value is within the preset range value; and b2”) if not, the application content information is marked for the administrator to confirm.

5. An artificial intelligence management system for generating diagnostic structured data, comprising: identifying a text diagnostic report issued by a hospital and provided by a user, having multiple fixed diagnostic fields for filling in multiple diagnostic information; and providing a structured data form required by an administrator, the structured data form having multiple information fields; wherein the artificial intelligence management system includes: A user-end processing device includes: an image capturing unit for capturing the surface image content of the user's text diagnostic report to generate diagnostic image data; and a user-end transmission unit for transmitting the diagnostic image data; an application interface server; an application interface terminal transmission unit for signal-connected to the user-end transmission unit and for receiving the diagnostic image data; and an application interface terminal processing unit including: a text recognition module for performing text recognition on the diagnostic image data; and a string conversion module for generating a set of text string information corresponding to the user based on the text recognition content of the text recognition module, using the fixed diagnostic fields and the diagnostic content; and a management processing server including: a management terminal transmission unit for signal-connected to the application interface terminal transmission unit and for receiving the text string information; A management processing unit includes an object conversion module for inputting diagnostic information corresponding to fixed diagnostic fields into the information fields of a structured data form based on the text string information, thereby generating a form file specific to the user; and a management storage unit for storing the form file; wherein the structured data form includes multiple different categories of medical codes, and the application interface processing unit determines whether the user's application content information corresponds to one of the medical codes based on the fixed diagnostic fields and / or the diagnostic content. If so, it generates corresponding medical code information, which the string conversion module then automatically generates the corresponding text string information specific to the user based on the application content information and the medical code information; wherein the medical codes are International Classification of Diseases (ICDs) or National Health Insurance (NHI) surgical codes.

6. The artificial intelligence management system as described in claim 5, wherein the structured data form includes multiple different categories of medical codes, wherein the application interface processing unit determines whether the user's application content information corresponds to a portion of the medical codes based on the application content information and the medical code information, and if so, recommends a portion of the medical codes and sends it back to the user terminal processing device, wherein the user terminal processing device transmits the user's selection to the application interface server through the user terminal transmission unit based on the user's selection of one of the recommended medical codes, so that the application interface processing unit generates corresponding medical code information based on the medical code selected by the user, and the string conversion module automatically generates the corresponding text string information specific to the user based on the application content information and the medical code information.

7. The artificial intelligence management system as described in claim 5, wherein the application interface processing unit further includes an intelligent learning module, and the management processing server further includes a management operation interface unit. When the administrator examines the structured data form electronic document and finds an error, the management operation interface unit receives a correction content input by the administrator and transmits the correction content to the application interface server through the management transmission unit. The intelligent learning module learns and corrects the identification and analysis of the application interface processing unit according to the correction instruction.

8. The artificial intelligence management system as described in claim 5, wherein the application interface processing unit includes an anomaly judgment module, and at least one of the information fields has a preset range value, wherein when the generated application content information has a value and the corresponding information field has the preset value, the application interface processing unit determines whether the value is within the preset range value, and if not, marks the application content information for the administrator to confirm.

Images