WO2010001792A1 - Database system - Google Patents

Abstract

Provided is a technique which can accurately identify an element constituting information.  A database system includes: a model storage unit (116) for storing model information; a structurizing unit (112) which decomposes input information into a plurality of elements according to the model information and adds item information to each of the elements so as to generate related information which correlates the plurality of elements; and a model update unit (116) which updates the model information in accordance with one or more elements constituting the input information for update and having an addition state of a predetermined identifier in a predetermined state, and item information relating to the one or more elements.

Description

Database system

The present invention relates to a database system.

When creating various reports such as reports using knowledge accumulated in the past, it has been conventionally necessary to search past reports by some method and refer to the contents. When past reports are structured using templates, etc., it is not necessary to refer to the full text of searched reports, but it was necessary to find out the relationship between reports while referring to a huge amount of data. .

And, for example, when creating a so-called interpretation report in a medical institution, if an input is made in a free format, the use of the abbreviations and phrases vary greatly depending on the individual physician of the radiology department. Therefore, it is difficult for other doctors to read, and when creating an interpretation report for the same case, it is difficult to create an interpretation report with reference to an interpretation report created by another radiology doctor.

For such a problem, in the list of information that supports the creation of the report, the combination of elements required for the feature description and the finding description based on the past cases is presented according to the frequency of the past cases. A technique has been proposed in which a sentence is created according to the selection of an element for each item (for example, Patent Document 1).

In the technique of Patent Document 1, a combination of elements is extracted by a search according to the designation of an element by the user, and information displayed in the list is narrowed down. In addition, in order to enable a search according to the designation of an element, a process in which a sentence written in a natural language (natural sentence) is decomposed into elements and a meaning attribute (semantic attribute) is given to each element. Some structuring is done. This structuring is executed by element recognition (pattern recognition) based on an element appearance pattern using, for example, a known SVM (Suppor Vector Vector Machine). In pattern recognition, since elements are generally identified according to a model indicating an element appearance pattern, it is necessary to learn the model in advance using teacher data.

JP 2007-140859 A

However, in structuring using a model that is learned based on teacher data, it is possible to perform high-precision structuring on natural sentences that show a similar tendency to teacher data, but similar to teacher data. For natural sentences that are not, high-precision structuring cannot be performed. That is, the performance of pattern recognition depends on the teacher data used for learning. For example, if the number of data constituting the teacher data is small, it is difficult to accurately structure natural sentences other than a certain pattern, and even if the number of data constituting the teacher data is large, there is a bias in content. Therefore, it is difficult to accurately structure natural patterns having various patterns. The occurrence of such a defect is due to the difficulty in accurately identifying the elements that make up a natural sentence.

Also, such a problem is not limited to the structuring of natural sentences, but is generally common when structuring various information.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a technique capable of accurately identifying elements constituting information.

In order to solve the above-described problem, a database system according to a first aspect includes model storage means for storing model information, and input information is decomposed into a plurality of elements according to the model information, and item information is stored in each element. Structuring means for generating related information in which the plurality of elements are associated with each other, one or more elements constituting the input information for update having a predetermined identification information addition state in a predetermined state, and the one or more elements Updating means for updating the model information on the basis of the item information relating to.

A database system according to a second aspect is the database system according to the first aspect, wherein the structuring unit generates related information related to the update input information, and the update unit includes the update input. The model information is updated based on related information related to the information.

The database system according to the third aspect is the database system according to the first aspect, wherein the update input information includes related information in which item information is assigned to each element and a plurality of elements are associated.

A database system according to a fourth aspect is the database system according to the second aspect, wherein output information for visually outputting related information related to the update input information, and the update input information by the output means Correction information receiving means for receiving related information in which the related information is corrected in response to an operation by an information input person in a state in which the related information is output in a visible manner, and the updating means includes The model information is updated based on the related information received by the correction information receiving means.

A database system according to a fifth aspect is the database system according to any one of the first to fourth aspects, wherein the predetermined identification information is added to the input information in response to an operation by the information input person Is provided with information input means for generating and inputting the update input information.

A database system according to a sixth aspect is the database system according to the fifth aspect, wherein one or more elements respectively associated with each item based on a plurality of related information generated by the structuring means A construction means for constructing a database in which a large number of combination information indicating a combination of a plurality of elements constituted by the above is stored, and in response to a request from a predetermined information user, And an information providing means for providing all or part of the information.

A database system according to a seventh aspect is the database system according to the sixth aspect, wherein the predetermined information user includes an information input person related to the update input information.

A database system according to an eighth aspect is the database system according to the sixth or seventh aspect, wherein the information is received in response to a request from an information user different from the information input person related to the update input information. The information processing device further comprises permission means for permitting provision of information to the information user by the information providing means while storing billing information for the user.

A database system according to a ninth aspect is the database system according to the second aspect, comprising a server and an information input terminal, wherein the server stores the model storage means, the structuring means, and the update And the information input terminal visually outputs the related information related to the update input information, and the output means outputs the related information related to the update input information visually. A correction information receiving unit that receives related information in which the related information is corrected in accordance with a user operation, and the updating unit is based on the related information received by the correction information receiving unit. To update the model information.

A database system according to a tenth aspect is the database system according to any one of the first to third aspects, comprising a server and an information input terminal, wherein the server includes the model storage means, Structuring means, and updating means, and the information input terminal sets the addition state of the predetermined identification information to the input information in response to a user operation, so that the input for updating Information input means for generating and inputting information is provided.

A database system according to an eleventh aspect is the database system according to the ninth or tenth aspect, further comprising a predetermined information use terminal, wherein the server generates a plurality of associations generated by the structuring means Based on the information, construction means for constructing a database in which a large number of combination information indicating combinations of a plurality of elements each composed of one or more elements associated with each item are stored, and the predetermined information utilization terminal And an information transmission means for transmitting all or part of the information in the database to the predetermined information use terminal in response to the request.

A database system according to a twelfth aspect is the database system according to the eleventh aspect, wherein the predetermined information utilization terminal includes an information input terminal in which the update input information is generated.

A database system according to a thirteenth aspect is the database system according to the eleventh or twelfth aspect, wherein the server requests from an information use terminal different from the information input terminal from which the update input information is generated. In response to this, it further has permission means for permitting transmission of information to the information utilization terminal by the information transmission means while storing charging information for the information utilization terminal.

A database system according to a fourteenth aspect is the database system according to any one of the sixth to eighth and eleventh to thirteenth aspects, wherein the database includes information indicating the number of combinations of a plurality of elements. .

The database system according to the fifteenth aspect is a database system according to any one of the sixth to eighth and eleventh to fourteenth aspects, and the combination information is described in RDF.

A database system according to a sixteenth aspect is the database system according to any one of the first to fifteenth aspects, wherein the structuring means identifies item information to be given to each element using SVM. .

A database system according to a seventeenth aspect is the database system according to any one of the first to sixteenth aspects, wherein the input information includes interpretation report information, endoscope report information, ultrasound report information, an incident It includes one or more of report information, discharge summary information, daily report information, report information, business document information, and patent specification information.

In the database system according to any one of the first to seventeenth aspects, the model information for structuring the input information is updated based on the input information in which the addition state of the predetermined identification information is the predetermined state. Desirable model information can be generated based on input information selected to some extent, and as a result, elements constituting the information can be accurately identified.

Even in the database system according to any of the fourth and ninth aspects, the input information is presented to the input person as a result of the structuring process of the input information, the correction is accepted, and the model information is updated based on the input information. Therefore, it is possible to update the model information more reflecting the intention of the input person.

In the database system according to any of the fifth and tenth aspects, the input person intentionally updates the model information based on the input information in which the addition state of the predetermined identification information is set to the predetermined state. Since model information is updated based on input information that is confident in content, better model information can be generated.

In the database system according to any of the seventh and twelfth aspects, the information stored in the database can be used for the person who provided the information for the update of the model information. Providing information about updates can be promoted. That is, the content of the model information can be enhanced by enhancing the information used for generating the model information.

By using the database system according to any of the eighth and thirteenth aspects, it is possible to use information stored in the database for a fee for those who do not provide information for updating model information. Further, it is possible to further promote provision of information for updating model information.

FIG. 1 is a diagram showing an overview of a DB system according to an embodiment of the present invention. FIG. 2 is a block diagram relating to the configuration of the DB system according to the embodiment of the present invention. FIG. 3 is a block diagram showing a functional configuration of the DB system according to the embodiment of the present invention. FIG. 4 is a diagram illustrating the interpretation report information. FIG. 5 is a diagram for explaining the structuring of the interpretation report information. FIG. 6 is a diagram illustrating a data configuration of the support information DB. FIG. 7 is a diagram illustrating a usage setting template. FIG. 8 is a diagram illustrating a correction confirmation template. FIG. 9 is a diagram illustrating a correction confirmation template. FIG. 10 is a diagram illustrating an examination list screen. FIG. 11 is a diagram illustrating a report input screen. FIG. 12 is a diagram illustrating requested items. FIG. 13 is a diagram illustrating the detailed contents of the inspection. FIG. 14 is a diagram illustrating a list of frequent phrases. FIG. 15 is a diagram illustrating a list of summary phrases. FIG. 16 is a diagram illustrating an extraction condition determination template. FIG. 17 is a diagram illustrating an input support template. FIG. 18 is a diagram showing a data flow in the data center server. FIG. 19 is a flowchart showing an operation flow of the DB system. FIG. 20 is a flowchart showing an operation flow of the DB system. FIG. 21 is a flowchart showing an operation flow of the DB system.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<1. Overall configuration of database system>
FIG. 1 is a diagram showing an overview of a database (DB) system 1 according to an embodiment of the present invention. The DB system 1 includes a data center server 100 and terminals 300a to 300c. Then, the data center server 100 and the terminals 300a to 300c are connected to each other via a network line NTW such as the Internet line so that data can be transmitted and received. The data center server 100 is installed in a service company that provides predetermined information, for example, and the terminals 300a to 300c are installed in different hospitals.

Although not shown in the drawings, each hospital is provided with a PACS (Picture Image Archiving System and Communication System) that stores a large number of image data acquired by examinations such as CT and MRI. . Each terminal 300a to 300b stores information (hereinafter referred to as “examination list information”) indicating a list (hereinafter also referred to as “examination list”) in which information relating to a large number of examinations is listed. The request information from the doctor in charge and the attribute information indicating the patient and examination information are stored in association with the examination IDs listed in the examination list.

In this DB system 1, the examination list information is output visually at each of the terminals 300a to 300c, and the interpretation doctor specifies the examination to be interpreted (hereinafter referred to as "interpretation examination"), thereby enabling the PACS server. Thus, it is possible to acquire image data related to the interpretation target inspection from the image and create a so-called interpretation report. Information indicating the interpretation report (hereinafter referred to as “interpretation report information”) input in each of the terminals 300a to 300c is stored in each of the terminals 300a to 300c and is also transferred to and stored in the data center server 100. . Then, the data center server 100 constructs various information based on the interpretation report information, and appropriately provides information to the terminals 300a to 300c in response to requests from the terminals 300a to 300c.

FIG. 2 is a block diagram relating to the configuration of the DB system 1 according to the embodiment of the present invention. As shown in FIG. 2, the data center server 100 has a general computer configuration in which a CPU 110 and a memory 150 are connected to a bus line 101. Note that the memory 150 appropriately includes a storage unit such as a RAM, a ROM, and a hard disk. The bus line 101 is connected to the network line NTW via an interface (I / F) 160. Various functions and operations of the data center server 100 are realized by the CPU 110 operating according to a program stored in the memory 150.

The terminals 300a to 300c have the same configuration as each other, and have a general computer configuration in which the CPU 310 and the memory 350 are connected to the bus line 301, respectively. The memory 350 appropriately includes a storage unit such as a RAM, a ROM, and a hard disk. In the memory 350, examination list information is stored, and attribute information indicating information on request items, patients, and examinations is listed in the examination list. Stored in association with the inspection ID. The bus line 301 is connected to the network line NTW through an interface (I / F) 360, and is connected to the display unit 330 and the operation unit 340 through an interface (not shown) as appropriate. The display unit 330 is configured by, for example, a liquid crystal display or a CRT, and the operation unit 340 has a keyboard, a mouse, and the like, and receives input of various signals in response to operations by the user. Here, since the terminals 300a to 300c have the same configuration, the terminal 300a will be described as a representative example as appropriate below.

<2. Functional configuration of database system>
FIG. 3 is a block diagram showing a functional configuration of the DB system 1 according to the embodiment of the present invention.

<2-1. Data center server functional configuration>
In the data center server 100, the CPU 110 reads and executes the program stored in the memory 150, so that the report receiving unit 111, the structuring unit 112, the report determining unit 113, the information transmitting unit 114, and the information receiving unit 115 are mainly used. , A model update unit 116, an information generation unit 117, a user management unit 118, a query reception unit 119, a search unit 120, and a result transmission unit 121 are realized. Data generated by each function of the data center server 100 is stored in the memory 150 as appropriate.

The report receiving unit 111 receives interpretation report information transmitted from the terminals 300a to 300c as input information. FIG. 4 is a diagram illustrating the interpretation report information. In FIG. 4, the interpretation report information described in the XML (Extensible Markup Language) format is illustrated, but other description formats may be used. Here, as shown in FIG. 4, in the interpretation report information, in addition to the information related to the text of the observation sentence described in the natural sentence, examination information (information relating to examinations such as examination apparatuses and interpretation doctors) and patient information Various attribute information such as (name of patient, sex, age, etc.) is included.

The interpretation report information received by the report receiving unit 111 is provided with information (hereinafter referred to as “transmission source identification information”) that identifies a transmission source such as a terminal-specific ID, IP address, and user ID. . In addition, the received interpretation report information (hereinafter referred to as model update use flag value) to which the received interpretation report information is given predetermined identification information that permits the use of the interpretation report information for the update of the model information to be described later (hereinafter referred to as the interpretation report information). There are two types of interpretation report information (hereinafter referred to as “unusable interpretation report information”) to which predetermined identification information is not assigned. Hereinafter, the update interpretation report information and the use prohibition interpretation report information are also abbreviated as “interpretation report information” as appropriate.

Further, the report receiving unit 111 constructs a report database (report DB) 151 by storing the received interpretation report information in the memory 150. The report DB 151 is composed of, for example, an RDB (Relational Database), but may be composed of other rules. Note that the interpretation report information stored in the report DB 151 may be searched for from each of the terminals 300a to 300c, and desired interpretation report information may be appropriately acquired and displayed on the terminals 300a to 300c. In addition, the report receiving unit 111 transfers the received interpretation report information to the structuring unit 112.

The structuring unit 112 structures the received interpretation report information. In the structuring unit 112, data constituting an interpretation report is structured and described in RDF according to model information stored in advance in the model storage unit 152 in the memory 150 (hereinafter referred to as “single report structuring”). Data ") is constructed. The model information is model data indicating how each element constituting the finding sentence included in the interpretation report should be decomposed into element items. Examples of methods for generating model information include the following. .

For example, a learning corpus including a large amount of text data conforming to the format of a finding sentence included in an interpretation report (specifically, a sentence model) is read. The sentence model indicates, for example, the structure of the finding sentence such as imaging conditions → site → basic findings (feature-conclusion) → diagnosis (diagnosis-conclusion). Moreover, in the learning corpus, the names of the classification items (hereinafter referred to as “items” where appropriate) of each element constituting the sentence model of the finding sentence are tagged to each word, and the name of the element item ( The item name) includes, for example, imaging conditions, parts, basic findings, diagnosis, and the like. Then, by machine learning, a phrase is extracted from the learning corpus and stored for each item of the corresponding element, thereby generating a part of the model information. In addition, by using machine learning using so-called SVM (Support Vector Vector Machine), the learning corpus is decomposed into morphemes by morphological analysis, and for each morpheme, the morpheme itself, the morpheme part of speech, the morpheme utilization form, front and back (for example, two before and after) A part of the model information is generated by learning a pattern in which a morpheme belonging to a certain item appears (hereinafter referred to as an “appearance pattern”) using the morpheme information).

Here, the process of structuring the interpretation report information in the structuring unit 112 will be briefly described. In the structuring unit 112, according to the model information, element items and words / phrases actually used are identified for the information of the observation sentence described in the natural sentence. For example, as shown in FIG. 5, various pieces of information included in one finding sentence are broken down into words (that is, elements) belonging to each element item. A plurality of elements are associated with each element while giving information indicating an item (hereinafter referred to as “item information”). At this time, elements indicating attribute information related to the interpretation report and attribute information related to examinations and patients, predetermined identification information, and elements indicating transmission source identification information (for example, ID and IP address unique to the terminal) are also appropriately included. In addition, the elements constituting one interpretation report are structured to construct single report structured data described in RDF. That is, the single report structured data related to the update interpretation report information includes predetermined identification information, and the single report structured data related to the use prohibited interpretation report information does not include predetermined identification information.

As a result of this structuring process, the item information corresponding to the finding sentence in the single report structured data is tagged with item information for each word / phrase in the same manner as the text data included in the learning corpus. It has become. The single report structured data includes information (that is, related information) in which a plurality of elements constituting a finding sentence included in the interpretation report information are associated with item information added to each element.

The report discriminating unit 113 discriminates whether or not the single report structured data structured and generated by the structuring unit 112 is used for updating the model information. Specifically, the report determination unit 113 determines that the single report structured data is used for updating the model information if predetermined identification information (here, model update use flag value) is included. If the predetermined identification information is not included, it is determined that the information is not used for updating the model information. Here, the single report structured data related to the use-banned interpretation report information is stored in the memory 150, thereby forming a related information database (related information DB) 153 in which the single report structured data is sequentially accumulated. The On the other hand, the single report structured data related to the update interpretation report information is transferred to the information transmission unit 114. Note that the single report structured data stored in the related information DB 153 can be searched for from the terminals 300a to 300c, and the desired single report structured data can be acquired and displayed as appropriate in the terminals 300a to 300c. Also good.

The information transmitting unit 114 receives the single report structured data generated by the structuring unit 112 for the interpretation report information to which the predetermined identification information is assigned, and the interpretation report information based on the single report structured data is input. To the terminal. Here, the transmission destination identification information included in the single report structured data is referred to identify the transmission destination terminal. For example, when the IP address is included in the transmission source identification information, the transmission destination terminal is recognized by the IP address.

The information receiving unit 115 receives single report structured data (hereinafter referred to as “determined single report structured data”) that has been corrected or confirmed by the terminal and whose contents are confirmed. The confirmed single report structured data received here is transferred to the model update unit 116. In addition, transmission source identification information is added to the fixed single report structured data, and this transmission source identification information is sent to the user management unit 118.

The model update unit 116 updates the model information stored in the model storage unit 152 based on the confirmed single report structured data transferred from the information reception unit 115. In the model update unit 116, the model information is updated by referring to one or more elements constituting the fixed single report structured data and item information that is an attribute assigned to each element.

Here, the update of model information (hereinafter also referred to as “model update”) will be described with a specific example. In the fixed single report structured data, the item information is tagged to each word / phrase in the same manner as the text data included in the learning corpus at the portion corresponding to the finding sentence. For this reason, in the model update unit 116, a part of the model information is updated by extracting a phrase from the fixed single report structured data and adding each of the corresponding element items to the model information. The Further, with respect to the fixed single report structured data, the appearance pattern of the element is recognized, and the appearance pattern is added to the model information, whereby a part of the model information is updated.

Also, the model update unit 116 adds the fixed single report structured data to the related information DB 153 and transfers it to the information generation unit 117.

The information generation unit 117 accumulates information indicating the structure of past interpretation report information as knowledge in the memory 150 using the confirmed single report structured data, thereby supporting information for supporting input of a new interpretation report. DB155 is generated. Specifically, a plurality of elements (in this case, attribute values) associated with each confirmed single report structured data sequentially input to the information generation unit 117 are arranged for each attribute item and re-associated with each other, and the RDF is used. Perform the process described. In other words, the support information DB 155 stores a large amount of combination information indicating combinations of a plurality of elements configured by one or more elements respectively associated with each item based on a plurality of fixed single report structured data in the memory 150. Is formed.

FIG. 6 is a diagram illustrating a data configuration included in the support information DB 155. As shown in FIG. 6, the support information DB 155 lists one or more attribute values (that is, elements, here, phrases) for each attribute item included in the plurality of attribute items, and between each attribute item, Information in which attribute values (that is, elements) are associated with each other in a network form (hereinafter referred to as “network information”) is included. In addition, when generating the network information, the information generation unit 117 counts the number of combinations of words and phrases between the items associated in each fixed single report structured data, and uses the count information as each support information. Store in the DB 155. For example, the number of combinations of phrases such as “head-MR-M-T1-frontal lobe-high signal-tumor” is stored in the count information. The network information and the count information are information indicating an association between one or more words belonging to each item among the items, and are hereinafter collectively referred to as “related element network information” as appropriate. Note that the count information may include information indicating the appearance frequency of the phrase.

The user management unit 118 stores the transmission source identification information input from the information reception unit 115 in the memory 150 as information for identifying the user (in this case, the terminal) that provided the information for updating the model information. Build a part of DB154. Also, the user management unit 118 receives a query indicating a search request for the support information DB 155 input from the terminals 300a to 300c via the query reception unit 119, and generates the query added to the query. Information (hereinafter referred to as “query generation source identification information”. For example, if the ID or IP address unique to the terminal) matches the transmission source identification information included in the user information DB 154, the query is simply Is transmitted to the search unit 120, and the query generation source identification information is transmitted to the result transmission unit 121.

On the other hand, when the query generation source identification information and the transmission source identification information included in the user information DB 154 do not match, the user management unit 118 charges the query generation source identification information and billing information (for example, the number of paid searches). ) Are stored in the user information DB 154. At this time, the query is sent to the search unit 120 and the query generation source identification information is sent to the result transmission unit 121. In other words, a terminal (hereinafter referred to as an “information utilization terminal”) that intends to use information by the search for the support information DB 155 by the user management unit 118 (hereinafter referred to as “information utilization terminal”) inputs the update interpretation report information (hereinafter referred to as “the information utilization terminal”). In the case of a search request from the information utilization terminal, the charging information for the information utilization terminal is stored in the memory 150, and the retrieval processing by the retrieval unit 120, Transmission of information indicating the search result by the result transmitting unit 121 to the information using terminal is permitted.

In the user information DB 154, a table in which source identification information for identifying users (in this case, terminals) that provided information for model update is listed, and query generation source identification information and accounting information for information search are associated with each other. May be stored separately, or information of two contents may be included in one table and stored.

The query receiving unit 119 receives queries generated by the terminals 300a to 300c. Query generation source specific information is given to this query.

The search unit 120 performs search processing for the support information DB 155 based on the query input from the user management unit 118, and sends a search result corresponding to the query to the result transmission unit 121. Here, the query includes, for example, an extraction condition indicating a combination of a plurality of elements, information related to a search result display format, and the like. Then, the search unit 120 extracts from the related element network information included in the support information DB 155 network information composed of combination information including combinations of elements that match the extraction conditions and count information related to the network information. The information extracted here (hereinafter referred to as “search result information”) corresponds to a part of the related element network information. Further, the search unit 120 generates a screen (result display screen) that generates information in which one or more elements are associated with each item based on the display format information of the search result included in the query, and displays the search result. Display data to be displayed (hereinafter referred to as “result display screen data”) is generated. The result display screen data is sent to the result transmission unit 121 together with the extracted count information (hereinafter also referred to as “extraction count information”).

The result transmission unit 121 transmits the result display screen data and the extraction count information input from the search unit 120 to the query transmission source terminal based on the query generation source identification information. Here, for example, when the IP address is included in the query generation source identification information, the destination terminal is recognized by the IP address. In this way, in response to a request for search processing from the terminal, a part of information constituting the support information DB 155 is extracted by the search processing in the search unit 120, and the extracted information is subjected to search processing by the result transmission unit 121. Sent to the requesting terminal.

<2-2. Functional configuration of terminal>
In the terminals 300a to 300c, the CPU 310 reads and executes the program stored in the memory 350, so that the report generation unit 311, the usage setting unit 312, the report transmission unit 313, the information reception unit 314, and the display control unit 315 are mainly used. , An information correction unit 316, an information transmission unit 317, a query generation unit 318, a query transmission unit 319, and a result reception unit 320 are realized. The terminals 300a to 300c function as terminals for inputting information (hereinafter referred to as “information input terminals”) when the interpretation doctor who is the user inputs the interpretation report information, and the interpretation doctor stores the information in the support information DB 155. When this information is used, it functions as a terminal (hereinafter referred to as “information use terminal”) that uses the information. Note that data generated by the functions of the terminals 300a to 300c is stored in the memory 350 as appropriate. Further, since the terminals 300a to 300c have the same configuration, the terminal 300a will be described as a representative example as appropriate.

The report generation unit 311 generates interpretation report information in response to the operation of the operation unit 340 by the interpretation doctor, and sends the interpretation report information to the use setting unit 312 as input information. When the interpretation report information is generated, the interpretation doctor becomes a person who inputs information (hereinafter referred to as “information input person”). In the terminal 300a, when inputting the interpretation report information, the interpretation doctor displays a part of the knowledge indicating the structure of the past interpretation report information stored in the support information DB 155 on the display unit 330 while interpreting the interpretation. Report information can be entered. The use of knowledge stored in the support information DB 155 will be described later.

In response to a predetermined operation of the operation unit 340 by the interpretation doctor, the usage setting unit 312 adds predetermined identification information (here, a model update usage flag value) to the interpretation report information, thereby updating the interpretation. Report information is generated and sent to the report transmission unit 313. In addition, the use setting unit 312 responds to other predetermined operations by the interpretation doctor, and adds the interpretation report information to the use prohibited interpretation report without adding predetermined identification information to the interpretation report information. The information is sent to the report transmission unit 313 as information.

FIG. 7 is a diagram exemplifying a template (usage setting template) T0 for setting whether or not to add predetermined identification information permitting use for model update to the interpretation report information when the interpretation report information is input. is there. In the usage setting template T0, two buttons BT11 and BT12 are provided, and an interpreting doctor appropriately operates a mouse included in the operation unit 340, so that the mouse pointer MP is aligned with the button BT11 and the left button of the mouse is pressed (left When the button is clicked, the button BT11 is pressed and predetermined identification information is added to the interpretation report information. FIG. 4 shows a state in which “use-model-updata =“ true ”” is described as predetermined identification information in the partial UI of the second line of the XML document. On the other hand, when the interpretation doctor operates the mouse as appropriate and the button BT12 is pressed, the predetermined identification information is not added to the interpretation report information. For example, the “use-model- updata = "false" "is described, and the use of the interpretation report information for model update is prohibited.

In addition, as shown in FIG. 7, the user does not set whether or not to add identification information manually, but depends on the attribute of the person, for example, years of experience, whether or not he / she is a resident, specialized area, etc. The system may determine semi-automatically the use of the interpretation report information for model update.

The report transmission unit 313 transmits the update interpretation report information and the use prohibition interpretation report information to the data center server 100 (specifically, the report reception unit 111). In other words, in response to the operation of the interpreting doctor, the report transmitting unit 313 receives the interpretation report information from the terminal (for example, the terminal 300a) to the data center server in the state of the updated interpretation report information to which predetermined identification information is appropriately added. 100 is input.

The information receiving unit 314 receives the single report structured data generated by the structuring unit 112 from the information transmitting unit 114 and outputs the data to the display control unit 315.

Based on the single report structured data input from the information receiving unit 314, the display control unit 315 displays information in which item information is added to each element constituting the finding sentence (that is, related information) in the display unit 330. Control to output visually. Here, the related information is output visually, and a template for accepting correction and confirmation of the related information (hereinafter referred to as “correction confirmation template”) is displayed on the display unit 330.

FIG. 8 is a diagram illustrating a correction confirmation template T1. In the correction confirmation template T1, a sentence (specifically, a finding sentence) registered in the upper part is described, related information is visually displayed in the central part, and three buttons BT21, BT22, and BT23 are displayed in the lower part. The

In the correction confirmation template T1 shown in FIG. 8, the registered finding sentence “A granular shadow is observed in the upper left lung field in the front image. Suspected inflammatory change.” Is displayed. The elements (phrases) belonging to the four items “imaging conditions”, “parts”, “basic findings”, and “diagnosis” constituting the finding sentence are described in the text boxes Bx1 to Bx4, respectively, and the prefix and suffix for each element Detailed elements (hereinafter referred to as “detailed element information”) belonging to detailed item information (hereinafter referred to as “detailed item information”) are appropriately described in a text box (for example, text box Bx21) that appears as necessary. It is displayed in the state. Specifically, the element “at the front image” belonging to the item information “imaging condition”, the element “lung field” belonging to the item information “part”, the element “granular shadow” belonging to the item information “basic findings”, and the item The element “inflammatory change” belonging to the information “diagnosis” is described. Further, the detailed element “upper left” belonging to the detailed item information “prefix” is described for the element “lung field”.

As described above, in the correction confirmation template T1, related information that is a result of structuring by the structuring unit 112 with respect to the information of the observation text included in the update interpretation report information is displayed on the display unit 330. In the correction confirmation template T1, the related information can be corrected. Specifically, the mouse pointer MP is moved to the text box whose description is to be corrected and left-clicked to cause the cursor to appear, and the text information is directly input using the keyboard included in the operation unit 340. You can correct the description. If you want to add a new word, that is, add an element that belongs to each item, press the add button with the mouse pointer MP to make the text box appear as appropriate and describe the desired element. Can do.

The information correction unit 316 corrects the related information in response to the operation of the operation unit 340 of the interpretation doctor in a state where the correction confirmation template T1 is visually output on the display unit 330. In addition, the correction confirmation template T1 is updated by sending information indicating the correction contents to the display control unit 315.

FIG. 9 is a diagram exemplifying a correction confirmation template T2 after the description is corrected with respect to the correction confirmation template T1 shown in FIG. Here, the term “upper left lung field” relating to the item information “part” is correctly decomposed and structured into “upper lung field” and its prefix “left”, but the correction shown in FIG. In the fixed template T1, “lung field” and its prefix “upper left” are erroneously decomposed and structured. On the other hand, in the correction confirmation template T2 shown in FIG. 9, the element belonging to the item information “part” is changed from “lung field” to “upper lung field” in the correction confirmation template T1 shown in FIG. The state after the modification which changes the detail element which belongs to the detailed item information “prefix” concerning “from the upper left” to “the left” is shown.

In the correction confirmation template T2, when the button BT21 is pressed with the mouse pointer MP, the process returns to the correction confirmation template T1 before the correction is made. When the button BT22 is pressed with the mouse pointer MP, the contents described in the correction confirmation template T2 are confirmed. At this time, the single report structured data is corrected or confirmed based on the related information determined by the information correcting unit 316, and the determined single report structured data (confirmed single report structured data) is converted into the information transmitting unit. Sent to 317. When the button BT23 is pressed with the mouse pointer MP, the display of the correction confirmation template T2 is terminated, and the operation for correcting and confirming related information is terminated. That is, the use of the single report structured data for updating the model information is stopped.

The information transmission unit 317 transmits the fixed single report structured data input from the information correction unit 316 to the data center server 100 (specifically, the information reception unit 115). Therefore, in a state where the related information related to the update interpretation report information is visibly output on the display unit 330, the related information is corrected and confirmed according to the operation of the operation unit 340 by the interpretation doctor, or The related information is simply confirmed, and the confirmed single report structured data is received by the information receiving unit 115.

The query generation unit 318 generates a query for performing a search for the support information DB 155 according to the operation of the operation unit 340 by the interpretation doctor at the time of creating the interpretation report, and sends the query to the query transmission unit 319. By this query, the data center server 100 is requested to download information for supporting creation of a finding sentence that matches a desired condition from the support information DB 155 (that is, use of past knowledge). Note that in the query generation unit 318, for example, information that identifies the terminal such as an ID unique to the terminal or an IP address (that is, query generation source identification information) is added to the query.

The query transmission unit 319 transmits the query input from the query generation unit 318 to the data center server 100 (specifically, the query reception unit 119).

The result receiving unit 320 receives the result display screen data and the extraction count information from the data center server 100 (specifically, the result transmitting unit 121). When the result receiving unit 320 receives the result display screen data and the extraction count information, the display screen data and the extraction count information are stored in the memory 350, and the display unit 330 is controlled under the control of the display control unit 315. A screen related to the result display screen data is displayed.

<3. Procedures for generating queries, using past knowledge, and entering findings>
Here, generation of a query, use of past knowledge, and input of a finding sentence when generating interpretation report information will be described.

First, a screen (hereinafter referred to as “examination list screen”) that visually outputs examination list information stored in the memory 350 is displayed on the display unit 330 in accordance with the operation of the operation unit 340 by the interpretation doctor. FIG. 10 is a diagram illustrating an examination list screen G1. The examination list screen G1 displays a list of information relating to examinations (specifically, patient ID, patient name, date of birth, age, sex, state, examination ID, examination date, examination site, modality, number of images). Screen. In this examination list screen G1, it can be seen that the examination with the state “uninterpreted” is an examination for which no interpretation report has been created. In addition, in the examination list screen G1, the thick frame cursor CS1 is moved up and down in accordance with the operation of the operation unit 340 by the image interpretation doctor, and the determination button (for example, the return key) of the operation unit 340 is set in accordance with the desired examination. ) Is specified, the single examination surrounded by the thick frame cursor CS1 is specified as the examination to be interpreted.

As described above, when the examination to be interpreted is specified, the screen displayed on the display unit 330 transitions from the examination list screen G1 to the report input screen G2 (FIG. 11). At this time, the CPU 310 identifies a task relating to the creation of a new interpretation report corresponding to the interpretation target examination, and reads request items corresponding to the interpretation target examination and attribute information indicating the details of the examination from the memory 350.

FIG. 11 is a diagram illustrating a report input screen G2. The report input screen G2 is mainly composed of areas A1 to A10. The area A1 displays the contents (for example, FIG. 12) of the requested items related to the interpretation target inspection, and the area A2 displays the detailed contents (for example, FIG. 13) of the inspection related to the interpretation target inspection. In the areas A3 and A4, free opinions and other information are described by the interpretation doctor. A part of information acquired from the support information DB 155 is visibly provided in the area A5 in order to support the creation of the interpretation report. In the area A6, finding sentences constituting the interpretation report are described according to the operation of the interpretation doctor. A list of standard phrases (frequent phrases) frequently used in the finding sentence (for example, FIG. 14) is displayed in the area A7, and a fixed phrase (summary phrase) frequently used in the finding sentence is displayed in the area A8. Is displayed (for example, FIG. 15). A representative image is attached to the area A9 in accordance with the operation of the interpretation doctor. In the area A10, icons for inputting commands (here, buttons BT1 and BT2) are listed. The button BT1 is a button for confirming and newly registering the findings and the contents of the attached image constituting the interpretation report created on the report input screen G2, and the button BT2 is the interpretation report created on the report input screen G2. Is a button for clearing the findings and the contents of the attached image.

Next, an operation for inputting a finding sentence using a template displayed in the area A5 will be described.

FIG. 16 is a diagram illustrating a template (extraction condition determination template) T3 that is displayed in the area A5 and determines extraction conditions for extracting some combination information from the support information DB 155. In the extraction condition determination template T3, for each item information “part classification”, “modality”, and “gender” of all the attribute items in the network information in the support information DB 155, one phrase (option) is selected for each item. A combination of options can be specified as an extraction condition. Specifically, in the extraction condition determination template T3 shown in FIG. 16, one attribute value (option) can be selected from each of the lists PL1 to PL3 by the mouse pointer MP. When one option is selected for each item information and the condition determination button BT3 is pressed with the mouse pointer MP, an extraction condition for specifying a combination of options (phrases) for the three item information is set. It is determined. At this time, the query generation unit 318 adds information related to the display format of the search result to the determined extraction condition, and generates a query.

Based on the query generated in this way, the search unit 120 uses the network information included in the support information DB 155 to generate network information including combination information including a combination of a plurality of elements that match the extraction condition and the network information. Count information is extracted. Then, the search unit 120 generates result display screen data including information in which one or more elements are associated with each item, based on the display format information of the search result included in the query. The result display screen data is transmitted to the result receiving unit 320 via the result transmitting unit 121 together with the extracted count information (extracted count information).

For example, in the terminal 300a, a list (list display) of information in which one or more elements are associated with each item is visually output on the display unit 330 under the control of the display control unit 315 based on the result display screen data. Is done. That is, a list display is visually output on the display unit 330 based on the information extracted by the search unit 120. Specifically, a template (hereinafter also referred to as “input support template”) that supports input of a finding sentence of a new interpretation report is displayed on the display unit 330.

FIG. 17 is a diagram showing a display example of the input support template T4. Note that various inputs and designations in the input support template T4 are performed based on signals input in response to an operation of the operation unit 340 by an interpreting doctor. As shown in FIG. 17, in the input support template T4, in the region T11 about 3/4 from the top, in order from the left, a plurality of elements (here, “imaging condition”, “part”, “basic findings”, “diagnosis”) Then, the phrases F1 to F4 are displayed. More specifically, in the input support template T4, a plurality of words (such as “front image” and “side image”) F1 for the attribute item “imaging condition” and a plurality of words (“lung field” for the attribute item “part”). F2 for the attribute item “basic findings” (for example, “total”, “apex apex”, etc.), and multiple words (for “pneumonia”, etc.) for the attribute item “diagnostic” for F3 F4 is listed.

In addition, for a plurality of elements listed in the input support template T4, one or more belonging to each item included in a plurality of items (here, four items “imaging condition”, “part”, “basic findings”, “disease name”) Display elements indicating the elements (here, phrases) are enumerated in spatial order. For example, as shown in FIG. 17, a display element indicating one or more words is listed under a display element indicating an item on which the one or more words are dependent, and a display element indicating a corresponding item. Are listed in a visually recognizable state. Specifically, for example, display elements “front image” and “side image” indicating a phrase are displayed side by side along the vertical direction below the display element indicating the item “shooting condition” as a subordinate destination. .

Then, in the input support template T4, for each of the four items “imaging condition”, “part”, “basic findings”, and “diagnosis”, the mouse pointer MP is moved from a plurality of words to at least one word and left-clicked. In addition, it is possible to specify words and phrases corresponding to each item constituting the finding sentence of the new interpretation report.

In addition, when an interpreting doctor who is a user designates a phrase corresponding to each item, the interpreting doctor refers to the sentence model MD of the finding sentence of the interpretation report displayed in the area T12 below the input support template T4. The information related to the sentence model MD may be given from the outside in advance and stored in the memory 350, for example. This sentence model MD indicates a model of an interpretation sentence of an interpretation report such as “[blank field W1] accepts [blank field W3] in [blank field W2]. Suspected of [blank field W4]”].

In the blank columns W1 to W4, words and phrases that are elements (elements constituting a sentence, hereinafter also referred to as “sentence constituent elements”) related to four items “imaging conditions”, “parts”, “basic findings”, and “diagnosis”. Each is true. That is, at least one word (for example, “front image”) of the plurality of words (options) F1 is applied to the blank column W1 by the designation of the words by the interpretation doctor, and at least one of the plurality of words (options) F2 is selected. One word (for example, “medium lung field”) is applied to the blank column W2. Furthermore, at least one of the plurality of phrases (options) F3 (for example, “equal shadow”) is applied to the blank field W3, and at least one of the plurality of phrases (options) F4 (for example, “lung cancer”). ) Is applied to the blank field W4. In addition, arbitrary words and phrases are applied to the blank columns W1 to W4 by various operations performed by the image interpretation doctor on the keyboard of the operation unit 340 and the like. In FIG. 17, four words “front image”, “medium lung field”, “equal shadow”, and “lung cancer” are designated for the four items “imaging condition”, “part”, “basic findings”, and “diagnosis”, respectively. A state in which the shade of the display element indicating the written phrase is inverted is schematically shown.

In such an input support template T4, as described above, when the words specified for each of the four items are applied to the blank columns W1 to W4 of the sentence model MD, the report generation unit 311 causes the report generation unit 311 to execute the blank column W1 of the sentence model MD. Words applied to .about.W4 are input as sentence constituent elements related to each item along a predetermined sentence model MD. At this time, a report sentence is generated by the report generation unit 311 based on the words / phrases assigned to the blank columns W1 to W4 of the sentence model MD and the sentence model MD and displayed in the area A5. When the operation unit 340 is appropriately operated and the button BT1 is pressed by the mouse pointer MP, the finding sentence displayed in the area A5 is determined as a new finding sentence, and a new sentence including the new finding sentence is included. An interpretation report is created. At this time, information of the new interpretation report is stored in the memory 150 in association with the examination list information, and is sent to the usage setting unit 312 so that the usage setting template T0 shown in FIG. 7 is displayed in the area A5. Then, in response to the input of the interpretation doctor for the usage setting template T0, the usage setting unit 312 generates update interpretation report information or usage prohibition report information, and the data center server 100 (specifically, via the report transmission unit 313). Is transmitted to the report receiving unit 111).

In the input support template T4 shown in FIG. 17, the element relation between the items is not shown in the network information, but is not limited to this. For example, the element relation between the items is two elements. It may be indicated by a display element such as a straight line connecting the two. Further, based on the extracted count information received by the result receiving unit 320 together with the result display screen data, the display element may be changed such that the thickness of a straight line connecting the two elements changes. For example, a combination of elements whose number of combinations is greater than or equal to a predetermined number in the extraction count information is displayed with a thick line, and a combination of elements whose number of combinations is less than the predetermined number in the extraction count information is displayed with a thin line. Such an embodiment is conceivable. In addition, when a certain phrase is specified, a straight line indicating a combination of the phrase and a phrase with a large number of combinations for the phrase may be emphasized by a thick line or the like. In the input support template T4 shown in FIG. 17, an example is given in which one sentence component is specified for each item. However, the present invention is not limited to this. For example, two or more sentence components are specified for each item. It may be configured such that at least one sentence component is specified, such as specifying.

<4. Data flow in the data center server>
FIG. 18 is a diagram illustrating a data flow in the data center server 100. First, when the interpretation report information D1 is received, it is stored in the report DB 151, and the interpretation report information D1 is structured by the structuring function P1 of the structuring unit 112 to generate single report structured data D2. Next, the report discriminating function P2 of the report discriminating unit 113 classifies the single report structured data D2 into data used for model update and data not used. Here, the single report structured data D3 not used for model update is stored in the related information DB 153 as it is. On the other hand, the single report structured data D4 used for model update is transmitted to the terminal by the information transmission function P3 of the information transmission unit 114. Thereafter, the fixed single report structured data that has been corrected and fixed at the terminal or simply fixed is received by the information receiving function P4 of the information receiving unit 115, and the fixed single report structured data D5 is stored in the related information DB 153. Is done. Further, the model update function P5 of the model update unit 116 updates the model information 152D in the model storage unit 152 based on the confirmed single report structured data D5. The updated model information 152D is used in the structuring process performed by the structuring unit 112 thereafter. Further, the information generation function P6 of the information generation unit 117 generates knowledge D6 indicating the configuration of past interpretation report information using the confirmed single report structured data, and accumulates it in the support information DB 155.

<5. Operation flow in DB system>
19 to 21 are flowcharts showing the operation flow of the DB system 1. This operation flow is realized by executing a predetermined program in the CPU 110 of the data center server 100 and the CPU 310 of the terminals 300a to 300c. Here, an example in which various processes are performed in the terminal 300a and the data center server 100 will be described.

In step S1, the examination list information in the memory 350 is read out in the terminal 300a, and the examination list screen G1 as shown in FIG. 10 is displayed on the display unit 330 under the control of the display control unit 315.

In step S2, it is determined by the determination function of the CPU 310 whether or not an inspection (interpretation inspection) that is an object to be interpreted on the inspection list screen G1 has been selected. Here, the determination in step S2 is repeated until the interpretation target inspection is selected. When the interpretation target inspection is selected, the process proceeds to step S3.

In step S 3, the report input screen G 2 as shown in FIG. 11 is displayed on the display unit 330 under the control of the display control unit 315.

In step S4, the extraction condition determination template T3 as shown in FIG. 16 is displayed in the area A5 of the report input screen G2 under the control of the display control unit 315.

In step S5, the query generation unit 318 generates a query including the extraction condition in response to the specification of the option on the extraction condition determination template T3, and the query is transmitted to the data center server 100 (query) via the query transmission unit 319. To the receiver 119). Note that query generation source identification information for identifying a terminal is added to the query.

In step S6, the user management unit 118 determines whether or not charging is required for the search according to the query generated in step S5. Here, it is determined whether or not charging is unnecessary by obtaining a query and query generation source identification information attached to the query from the query receiving unit 119 and referring to the user information DB 154. Specifically, when the query generation source identification information matches the transmission source identification information included in the user information DB 154, it is determined that charging is unnecessary, the query is sent to the search unit 120, and the query The generation source identification information is sent to the result transmission unit 121, and the process proceeds to step S8. On the other hand, if the query generation source identification information does not match the transmission source identification information included in the user information DB 154, it is determined that charging is necessary, and the process proceeds to step S7.

In step S7, the user management unit 118 associates the billing information for the information search with the query generation source identification information and stores it in the user information DB 154. At this time, the query is sent to the search unit 120 and the query generation source identification information is sent to the result transmission unit 121.

In step S8, the search unit 120 extracts some information (specifically, network information and count information) corresponding to the extraction conditions included in the query generated in step S5 from the support information DB 155. . Here, the search unit 120 generates result display screen data based on the extracted part of the network information, and transmits the result display screen data to the result reception unit 320 via the result transmission unit 121 together with the extracted count information.

In step S9, the display support unit 315 displays the input support template T4 on the display unit 330 based on the result display screen data generated in step S8.

In step S10, the report generation unit 311 accepts an input from the operation unit 340 by the interpretation doctor, and performs various inputs in the input support template T4. At this time, a finding sentence corresponding to the designation of the phrase in the input support template T4 is presented in the area A6.

In step S11, the report generator 311 determines whether or not there is an instruction to register the interpretation report information. Here, various inputs in the input support template T4 are accepted until the button BT1 is pressed on the report input screen G2 and an instruction to register interpretation report information is given (step S10), and an instruction to register interpretation report information is issued. If done, the process proceeds to step S21 in FIG.

In step S21 of FIG. 20, the report generation unit 311 generates image interpretation report information. At this time, the interpretation report information is sent to the usage setting unit 312.

In step S22, the use setting template T0 shown in FIG. 7 is displayed on the display unit 330 under the control of the display control unit 315.

In step S23, it is determined whether or not to add predetermined identification information that permits the use of interpretation report information for model update. Here, in the usage setting template T0, if the button BT11 is pressed, it is determined that the predetermined identification information is added to the interpretation report information, and the process proceeds to step S24. If the button BT12 is pressed, It is determined that the predetermined identification information is not added to the interpretation report information, and the process proceeds to step S25.

In step S24, the use setting unit 312 adds predetermined identification information to the interpretation report information generated in step S21 to generate update interpretation report information.

In step S25, the use setting unit 312 does not add the predetermined identification information to the interpretation report information generated in step S21, and the use prohibition interpretation report information is generated.

In step S26, the report transmission unit 313 transmits the interpretation report information to the data center server 100 (specifically, the report reception unit 111). Specifically, when the process proceeds from step S24, the update interpretation report information generated at step S24 is transmitted to the report reception unit 111. When the process proceeds from step S25, the process proceeds to step S25. The generated prohibited use interpretation report information is transmitted to the report receiving unit 111.

In step S27, the report receiving unit 111 stores the interpretation report information transmitted in step S26 in the report DB 151. At this time, the interpretation report information is also sent to the structuring unit 112.

In step S28, the structuring unit 112 structures the information of the finding text included in the interpretation report information sent from the report receiving unit 111 in step S27.

In step S29, attribute information is added to the finding sentence information structured in step S28 by the structuring unit 112, and single report structured data is generated. When the interpretation report information to be structured is the interpretation report information for update, the single report structured data includes elements that include the specified identification information that permits the use of the interpretation report information for model update. Described.

In step S30, the report determination unit 113 determines whether or not the single report structured data generated in step S29 can be used for model update. Here, when the predetermined identification information is included in the single report structured data, it is determined that the single report structured data can be used for model update, and the process proceeds to step S31. On the other hand, when the predetermined identification information is not included in the single report structured data, it is determined that the single report structured data cannot be used for model update, and the process proceeds to step S32.

In step S31, the information transmission unit 114 transmits the single report structured data generated in step S29 to the terminal 300a (specifically, the information reception unit 314), and the process proceeds to step S41 in FIG.

In step S32, the single report structured data generated in step S29 is stored in the related information DB 153 by the structuring unit 112, and this operation flow ends.

21, under the control of the display control unit 315, the correction confirmation template T1 as shown in FIG. 8 is displayed on the display unit 330 based on the single report structured data transmitted in step S31. .

In step S42, the information correction unit 316 receives an input from the operation unit 340 by the interpretation doctor, and performs various inputs in the correction confirmation template T1.

In step S43, the information correction unit 316 determines whether or not use of the single report structured data for model update is permitted. Here, if the button BT22 is pressed in the correction confirmation template T1, it is determined that the use for model update is permitted, the contents of the single report structured data are confirmed, and the process proceeds to step S44. On the other hand, if the button BT23 is pressed in the correction confirmation template T1, it is determined that the use for model update is not permitted, and this operation flow ends.

In step S44, the information transmission unit 317 sends the single report structured data (confirmed single report structured data) whose contents are confirmed in step S43 to the data center server 100 (specifically, the information receiving unit 115). Sent.

In step S45, the model updating unit 116 performs storage processing for storing the confirmed single report structured data transmitted in step S44 in the related information DB 153.

In step S46, the model update unit 116 updates the model information in the model storage unit 152 based on the confirmed single report structured data transmitted in step S44.

In step S47, the user management unit 118 generates or updates the user information DB 154 based on the transmission source identification information included in the fixed single report structured data transmitted in step S44. Here, when the user information DB 154 is not constructed in the memory 150, the transmission source identification information is stored in the memory 150 as information for identifying the terminal that has provided the information for updating the model information. DB 154 is newly generated. On the other hand, when the user information DB 154 is stored in the memory 150, the source information DB 154 is updated by adding the transmission source specifying information to the information specifying the terminal that provided the information for updating the model information. Done.

In step S48, the information generation unit 117 generates or updates the support information DB 155 based on the confirmed single report structured data transmitted in step S44. Here, when the support information DB 155 is not constructed in the memory 150, the support information DB 155 is newly constructed. On the other hand, when the support information DB 155 is stored in the memory 150, the element-related network information included in the support information DB 155 is updated based on the new confirmed single report structured data. When the process of step S48 is completed, this operation flow is ended.

As described above, in the DB system 1 according to the embodiment of the present invention, the model information for structuring the interpretation report information including the information of the observation sentence described in the natural sentence is added with the predetermined identification information. Updated based on interpretation report information. Therefore, it is possible to generate preferable model information based on interpretation report information selected to some extent. As a result, it is possible to accurately identify elements constituting the information.

In addition, after presenting the result of the structured processing related to the interpretation report information to the interpreting doctor who is the input person of the interpretation report information, and accepting corrections to the result, updating the model information based on the interpretation report information Is done. For this reason, it is possible to update the model information more reflecting the intention of the input person, and the structuring by the biased model information is corrected.

Also, the model information is updated based on the interpretation report information to which predetermined identification information is intentionally added by the interpretation doctor who is the input person of the interpretation report information. For this reason, the model information is updated based on the interpretation report information that the interpretation doctor who is the input person is confident in content, and a better model can be generated.

In addition, the information stored in the support information DB 155 can be used for a person who provides information for updating model information. Furthermore, for those who do not provide information for updating model information, the information stored in the support information DB 155 can be used for a fee. For this reason, the incentive with respect to the provision of the information with respect to the update of model information can be given, and the provision of the information with respect to the update of the model information can be promoted. As a result, the content of the model information can be enhanced by enhancing the information used for generating the model information.

<6. Modification>
The present invention is not limited to the above-described embodiments, and various changes and improvements can be made without departing from the scope of the present invention.

For example, in the above-described embodiment, the interpretation report information used for updating the model information is input to the data center server 100 from the terminals 300a to 300c in a state including the information of the observation sentence described in the natural sentence. However, it is not limited to this. For example, the interpretation report information may be input from the terminals 300a to 300c to the data center server 100 in a structured state, that is, in a single report structured data state. Further, for example, in each of the terminals 300a to 300c, only the update interpretation report information may be structured and then transmitted to the data center server 100, or the update interpretation report information and the use prohibited interpretation report information may be transmitted. After at least one of them is structured, it may be transmitted to the data center server 100.

In the above embodiment, the predetermined identification information that permits use for model update is added to the interpretation report information. However, the present invention is not limited to this. For example, the use of interpretation report information for model update is permitted. In the case of prohibiting the use of interpretation report information for model update without adding any particular information to the interpretation report information, other predetermined identification information that prohibits use for model update in interpretation report information May be added. That is, the interpretation report information in which the addition state of the predetermined identification information is in the predetermined state may be used as update interpretation report information for which use of information is permitted for model update.

In the above embodiment, there are only three terminals (terminals 300a to 300c) that provide information to the data center server 100, but the present invention is not limited to this, and a plurality of terminals may be used. However, from the viewpoint of enhancing the contents of the model information, it is preferable that the number of terminals that provide information to the data center server 100 is large.

In the above embodiment, the network line NTW that connects the data center server 100 and the terminals 300a to 300c so as to be able to transmit and receive data is configured by the Internet line. However, the present invention is not limited to this. You may comprise by a communication line. For example, there may be a mode in which one data center server 100 is provided in a large hospital, a terminal is provided in each clinical department, each examination room, and the like, and is connected so as to be able to transmit and receive data via a LAN.

In the above embodiment, information is provided from a plurality of terminals, but the present invention is not limited to this. For example, when a single terminal is shared by a plurality of users, it is recognized as a different information provider or information user for each logged-in user by logging in using a user ID or password. Also good. That is, even when information is provided from the same terminal or when information is used, when the user who provided the interpretation report information for model update is using the terminal, the information in the support information DB 155 is free of charge. When a user who does not provide interpretation report information for model update is using a terminal, the information in the support information DB 155 may be available for a fee.

As specific examples, the following modes are conceivable. In the user information DB 154, a user who provides interpretation report information for model update and a user who does not provide the model update are distinguished and managed by a personal user ID or password. When the user who has input the update interpretation report information requests a search for the support information DB 155, the information is searched for free and the search result is provided to the terminal where the user is logged in. To do. On the other hand, when a user different from the user who has input the update interpretation report information requests a search for the support information DB 155, the charging information is associated with the user ID of the individual user who requested the search. While storing, the search result is provided to the terminal to which the user is logged in. Even in such an aspect, the user management unit 118 permits transmission (that is, provision) of search result screen data to the information user by the result transmission unit 121.

In the above embodiment, a so-called client-server system including the data center server 100 and a plurality of terminals 300a to 300c is used. However, the present invention is not limited to this. For example, one personal computer may have the functions of the data center server 100 in the DB system 1 and the functions of a plurality of terminals 300a to 300c. As a method of providing the functions of a plurality of terminals 300a to 300c in one personal computer, a method in which terminals logged in by different users are regarded as a plurality of different terminals by login using a user ID, a password, or the like. Is mentioned.

In the above embodiment, model update, construction of the related information DB 153, construction of the support information DB 155, and the like have been performed by taking as an example the case where interpretation report information is input. However, the present invention is not limited to this. For example, it is possible to apply the configuration of the present invention to input of various report information other than interpretation report information. Examples of report information other than the interpretation report information include, for example, report information (endoscope report information) indicating the results of inspection using an endoscope, and report information (super information) indicating the results of inspection using ultrasound. Sonic report information), report information describing incident contents, causes and countermeasures (incident report information), and various medical-related report information such as so-called discharge summary information. In addition, various report information such as daily report information describing daily events and situations, information on various reports (report information), and information on various business documents (business document information) can be considered in fields other than medical related fields. . Also, the configuration of the present invention can be applied to documents other than various report information, for example, patent specification information (patent specification information) as input targets. That is, the update interpretation report information may be input information for updating various documents (hereinafter also referred to as “update input information”).

In the above-described embodiment, the aspect in which a part of the related element network information included in the support information DB 155 is extracted and used is described. However, the present invention is not limited to this. For example, an aspect in which all of the related element network information included in the support information DB 155 is extracted and used in response to a user request is also conceivable.

In the above embodiment, the combination of a plurality of elements constituting the extraction condition is specified in the extraction condition determination template T3. However, the present invention is not limited to this. For example, when an interpretation target examination is selected on the examination list screen G1 as shown in FIG. 10, an element (eg, chest) belonging to the item “examination part” and an element (eg, CR) belonging to the item “modality” Combinations may be recognized at the same time and included in a combination of a plurality of elements constituting the extraction condition.

In the above embodiment, information in the support information DB 155 is provided for a fee to a user or terminal that does not provide information for model update, but information is not provided for model update. Information in the support information DB 155 may not be provided to the user or the terminal at all.

Note that, of course, the above-described modified examples may be appropriately combined as long as no contradiction occurs.

1 Database System 100 Data Center Server 110, 310 CPU
DESCRIPTION OF SYMBOLS 111 Report reception part 112 Structured part 113 Report discrimination | determination part 114 Information transmission part 115 Information reception part 116 Model update part 117 Information generation part 118 User management part 119 Query reception part 120 Search part 121 Result transmission part 150,350 Memory 151 Report database (Report DB)
152 Model Storage Unit 153 Related Information Database (Related Information DB)
154 User information database (user information DB)
155 Support Information Database (Support Information DB)
300a to 300c Terminal 311 Report generation unit 312 Usage setting unit 313 Report transmission unit 314 Information reception unit 315 Display control unit 316 Information correction unit 317 Information transmission unit 318 Query generation unit 319 Query transmission unit 320 Result reception unit 330 Display unit 340 Operation unit

Claims (17)

1. A database system,
   Model storage means for storing model information;
   In accordance with the model information, structuring means for generating related information that associates the plurality of elements while decomposing the input information into a plurality of elements and adding item information to each element;
   Updating means for updating the model information based on one or more elements constituting the update input information in which the addition state of the predetermined identification information is in the predetermined state and item information relating to the one or more elements;
   A database system comprising:
2. The database system according to claim 1, wherein
   The structuring means comprises:
   Generating related information related to the update input information;
   The updating means
   The database system, wherein the model information is updated based on related information related to the update input information.
3. The database system according to claim 1, wherein
   The update input information is
   A database system characterized in that item information is given to each element and includes related information in which a plurality of elements are associated.
4. The database system according to claim 2, wherein
   An output means for visually outputting related information relating to the update input information;
   In a state where the related information related to the update input information is visibly output by the output means, a correction information receiving means for receiving the related information in which the related information is corrected according to an operation by the information input person;
   Further comprising
   The updating means
   A database system, wherein the model information is updated based on related information received by the correction information receiving means.
5. A database system according to any one of claims 1 to 4, wherein
   An information input means for generating and inputting the update input information by setting the addition state of the predetermined identification information to the input information in a predetermined state in response to an operation of an information input person;
   A database system comprising:
6. The database system according to claim 5, wherein
   Based on a plurality of related information generated by the structuring means, a database is constructed in which a large number of combination information indicating combinations of a plurality of elements each composed of one or more elements associated with each item is stored. Construction means,
   Information providing means for providing all or part of the information of the database to the predetermined information user in response to a request from the predetermined information user;
   A database system further comprising:
7. The database system according to claim 6, wherein
   The predetermined information user
   A database system comprising an information input person related to the update input information.
8. The database system according to claim 6 or 7, wherein
   In response to a request from an information user different from the information input person related to the update input information, the information providing means permits the information user to provide information while storing the billing information for the information user Permission means,
   A database system further comprising:
9. The database system according to claim 2, wherein
   A server and an information input terminal,
   The server is
   The model storage means, the structuring means, and the updating means,
   The information input terminal is
   An output means for visually outputting related information relating to the update input information;
   In a state where the relevant information related to the update input information is visibly output by the output means, a correction information accepting means for accepting relevant information in which the relevant information is modified in response to a user operation;
   Have
   The updating means
   A database system, wherein the model information is updated based on related information received by the correction information receiving means.
10. The database system according to any one of claims 1 to 3, wherein
    A server and an information input terminal,
    The server is
    The model storage means, the structuring means, and the updating means,
    The information input terminal is
    In response to a user operation, an information input unit that generates and inputs the update input information by setting the addition state of the predetermined identification information to the input information to a predetermined state;
    A database system comprising:
11. The database system according to claim 9 or 10, wherein
    A predetermined information use terminal,
    Further comprising
    The server is
    Based on a plurality of related information generated by the structuring means, a database is constructed in which a large number of combination information indicating combinations of a plurality of elements each composed of one or more elements associated with each item is stored. Construction means,
    Information transmitting means for transmitting all or part of the information in the database to the predetermined information utilization terminal in response to a request from the predetermined information utilization terminal;
    A database system further comprising:
12. The database system according to claim 11, wherein
    The predetermined information using terminal is
    A database system comprising an information input terminal in which the update input information is generated.
13. A database system according to claim 11 or claim 12,
    The server is
    In response to a request from an information utilization terminal different from the information input terminal from which the update input information is generated, the information transmission means transmits information to the information utilization terminal while storing charging information for the information utilization terminal. Permission means to allow,
    A database system further comprising:
14. A database system according to any one of claims 6 to 8 and claim 11 to claim 13,
    The database is
    A database system comprising information indicating the number of combinations of a plurality of elements.
15. The database system according to any one of claims 6 to 8 and claim 11 to claim 14,
    The combination information is
    A database system characterized by being described in RDF.
16. The database system according to any one of claims 1 to 15, wherein
    The structuring means comprises:
    A database system characterized by identifying item information to be given to each element using SVM.
17. The database system according to any one of claims 1 to 16, comprising:
    The input information is
    Including one or more information of interpretation report information, endoscope report information, ultrasound report information, incident report information, discharge summary information, daily report information, report information, business document information, and patent specification information A featured database system.

Images