Drug Information Recognition and Alert System Based on a Non-Integrated Pharmacy Information System Screen

Abstract

The present invention relates to a drug data or screen-based relationship analysis and adaptive notification system, comprising: a data acquisition unit (50) configured to receive drug data including drug name, drug code, ingredient information, and content information, or to generate said drug data from a screen of a pharmacy computer system; a relationship analysis unit (60) that analyzes relationships between drugs based on said drug data, including whether ingredients match, whether they belong to similar classes, and relationships regarding contraindications or precautions for concomitant use between multiple drugs; a history-based adaptive risk judgment unit (70) that determines whether a notification is to be issued by combining the results of the relationship analysis with drug information or notification conditions of interest pre-set by a user, and dynamically controls whether a notification is to be issued or the method of outputting the notification by reflecting the previous detection history, detection frequency, and user notification response history of the same drug or drugs of the same ingredient group; and an adaptive notification output control unit (80) that outputs a notification including drug information and a reason for notification when a notification is required according to the result of the judgment, and dynamically adjusts the output method, frequency, or priority of the notification.

Description

Technology Field

[0001] The present invention relates to a drug data or screen-based relationship analysis and adaptive notification system. Background Technology

[0003] Recently, it has become common practice in pharmacies to perform prescription entry, drug information management, and dispensing tasks through computerized systems, and pharmacists perform dispensing and patient guidance based on the drug lists displayed on the screen. While these pharmacy computerized systems provide functions to display or manage drug names, codes, ingredient information, and dosage details, most are limited to simple information input and retrieval functions. Generally, features for analyzing drug relationships or providing automatic notifications based on specific conditions are either not included or are provided in a very limited manner.

[0004] Consequently, pharmacists must individually verify drug information displayed on the computer screen to determine whether there are duplicates of drugs with the same active ingredient, whether there are contraindications or precautions for concomitant use, and whether specific drugs require caution; this presents a problem in that such judgments rely heavily on the pharmacist's experience and proficiency. In particular, when multiple drugs are prescribed simultaneously, drug interactions must be considered comprehensively, so there are limitations in ensuring safety through simple list verification alone. Furthermore, the absence of a systematic notification function that automatically recognizes and informs the user when caution is required regarding specific drugs or drug combinations creates a possibility of errors or omissions during the dispensing process.

[0005] Furthermore, pharmacy computer systems often operate with a closed structure, frequently limiting integration with external systems or making data access difficult. Consequently, methods that directly receive and analyze drug data from external systems face limitations in practical application. While some technologies have proposed attempts to collect and analyze drug data by linking with external systems, these approaches are difficult to apply universally due to the varying structures of computer systems used by different pharmacies and the frequent lack of provided integration interfaces.

[0006] Meanwhile, to overcome these limitations, some technologies have been proposed that capture pharmacy computer screens and recognize drug information through OCR (Optical Character Recognition). However, conventional OCR-based technologies often remain at the level of simply extracting text displayed on the screen, and they have limitations in that they do not sufficiently consider functions for accurately structuring drug names, codes, ingredients, and dosage information, or for precisely analyzing the relationships between drugs based on this information.

[0007] Furthermore, in conventional technology, even when relationship analysis between drugs is performed, it often relies on simple name or code matching, which fails to adequately reflect the relationships between drugs with the same active ingredient or similar classes, and there are limitations in comprehensively analyzing combination relationships between multiple drugs, such as contraindications or precautions for concomitant use.

[0008] Furthermore, even if a notification function exists in conventional technology, it is limited to providing notifications uniformly only when specific conditions are met based on fixed criteria. There is a lack of customized features that reflect the drugs or notification conditions that users wish to manage as important, and there is a problem with insufficient functionality to dynamically control notifications by considering the repeated detection of the same drug or user response history.

[0009] In other words, conventional pharmacy computer environments fail to provide integrated technologies for flexibly acquiring drug data, precisely analyzing relationships between drugs, user-customized decision functions, and history-based adaptive notification control. In particular, there is a limitation in the absence of a systematic system that automatically analyzes drug relationships and provides timely notifications to users.

[0010] Therefore, there is a need to develop a technology that is applicable in various environments by including both methods of receiving drug data from external systems and methods of generating drug data from pharmacy computer system screens, while also precisely analyzing the relationships between multiple drugs and adaptively controlling notifications by reflecting user-configured conditions and past history. The problem to be solved

[0012] The present invention has been devised to solve the above-mentioned problems and provides a drug data or screen-based relationship analysis and adaptive notification system.

[0013] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0015] A drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention comprises: a data acquisition unit (50) configured to receive drug data including drug name, drug code, ingredient information, and content information, or to generate said drug data from a screen of a pharmacy computer system; a relationship analysis unit (60) that analyzes relationships between drugs including whether ingredients match, whether they belong to similar classes, and relationships of contraindications or precautions for concomitant use between a plurality of drugs based on said drug data; a history-based adaptive risk judgment unit (70) that determines whether a notification is to be issued by combining the results of the relationship analysis with drug information or notification conditions of interest pre-set by a user, and dynamically controls whether a notification is to be issued or the method of outputting the notification by reflecting the previous detection history, detection frequency, and user notification response history of the same drug or drugs of the same ingredient group; and an adaptive notification output control unit (80) that outputs a notification including drug information and a reason for notification when a notification is required according to the result of the judgment, and dynamically controls the method of outputting the notification, frequency, or priority.

[0016] A drug data or screen-based relationship analysis and adaptive notification system according to an embodiment of the present invention comprises: a screen capture unit (100) that periodically captures a display screen of a pharmacy computer system running on a local pharmacy computer or at the time of a preset event occurrence; a region identification unit (200) that identifies a region of interest in which drug information is displayed from a screen image obtained by the screen capture unit (100) using at least one of coordinate information, screen layout information, color pattern, character arrangement pattern, and arrangement form of user interface objects; a character recognition unit (300) that extracts characters, numbers, or symbols included in the region of interest using an optical character recognition method to generate at least one or more drug-related information among a drug name, drug code, content, dosage quantity, number of doses, and number of days of administration; and a data structuring unit (400) that converts the drug-related information extracted by the character recognition unit (300) into structured data in row units, column units, or item units, and aligns the characters located in the same row or adjacent rows into prescription data in individual drug units by associating them with each other. Error correction unit (500) that corrects misrecognition, omission, or separation recognition occurring during the optical character recognition process by utilizing the correspondence between the drug name, drug code, content, formulation, specification, dosage quantity, number of doses, and number of days of administration included in the above structured data, and improves the accuracy of drug identification by comparing it with standard drug information included in a stored drug database; drug matching unit (600) that determines the ingredient information, whether the ingredients are identical, contraindications for concomitant use, whether specific caution ingredients are included, whether it is a user-designated drug of interest, and whether it is a pre-set notification target by comparing the structured data that has passed through the error correction unit (500) with the stored drug database;The present invention is characterized by including a judgment and notification unit (700) that determines whether to generate a notification by applying a preset risk judgment rule or a user-defined notification rule according to the above drug matching result, and if a notification is required, outputs notification information including at least one of a drug name, drug code, dosage, dosage quantity, number of doses, and number of days of administration in the form of a screen, pop-up window, voice, color highlighting, or message.

[0017] In one embodiment, the bending device may include: a bottom support member formed in a rectangular shape that supports a floor at the bottom and a bending member at the top; a workbench installed longitudinally on the upper surface of the bottom support member to provide a bending workspace; a mold member installed longitudinally on the upper surface of the workbench to support and place a workpiece; and a bending member installed parallel to the longitudinal direction on the upper side of the mold member to apply pressure to bend the workpiece.

[0018] In one embodiment, the screen capture unit (100) captures at least one of the prescription input screen, dispensing screen, medication guidance screen, payment screen, or drug list display screen of a pharmacy computer system, and the preset event includes at least one of the completion of prescription QR recognition, prescription data retrieval, screen switching, cursor position change, drug list update, change of dispensing patient, input of a specific shortcut key, mouse click, change of string within the screen, change of a specific color area, or activation of a specific UI object, and the screen capture unit (100) may be configured to perform a subsequent recognition procedure only when it is confirmed that the drug list has changed by comparing the screen before and after the occurrence of the event.

[0019] In one embodiment, the area identification unit (200) distinguishes a table area, list area, cell area, column title area, quantity display area, and total display area where drug information is displayed within the screen, and the character recognition unit (300) applies different character recognition conditions to each identified area, wherein a Korean priority recognition condition is applied to the Korean drug name area, a number priority recognition condition is applied to the drug code area, and a mixed recognition condition of numbers and unit strings is applied to the content and specification area, respectively. The data structuring unit (400) groups adjacent text blocks on the same horizontal line or within a preset error range into a single drug row, stores the extracted drug name, drug code, content, quantity, number of doses, and number of days of administration for each row by corresponding them to attribute fields, and is configured to individually generate a structured record for each drug when multiple drugs are listed.

[0020] In one embodiment, the error correction unit (500) is characterized by performing a cross-comparison based on standard drug name, drug code, ingredient name, dosage form, content, and specification information included in a stored drug database when either the drug name or the drug code extracted as a result of optical character recognition is incomplete or misrecognized, deriving a correction candidate using a number sequence similarity and a preset digit pattern when only some digits of the drug code are recognized or replaced with similar numbers, deriving a correction candidate using a string similarity, a stored drug dictionary, preset prohibited words, and a drug name prefix or suffix pattern when the drug name is separated into some syllable units or misrecognized with similar characters, deriving a correction candidate using a string similarity, a stored drug dictionary, preset prohibited words, and a drug name prefix or suffix pattern, confirming a candidate among the correction candidates whose degree of agreement with the drug name, drug code, content, and dosage form is greater than or equal to a reference value as the final drug information, and when the drug name and drug code indicate different drugs, using the drug code as the priority standard or selectively applying one of the drug name priority standard, code priority standard, or multiple candidate warning output standard according to user settings. You can do it.

[0021] In one embodiment, the drug matching unit (600) and the judgment and notification unit (700) determine whether to generate a notification based on at least one of the following criteria, based on drugs included in a stored drug database or a user-registered notification list: whether there is a duplicate prescription of the same ingredient, whether there is a pre-set specific ingredient, whether there is a combination of contraindications or precautions for concomitant use, whether there is a duplicate of drugs of the same class, whether there is a drug of caution for a specific age group or a specific disease group, whether there is an excess of a pre-set dosage or pre-set quantity, whether there is a drug of interest registered in advance by the user, and whether there is a match with the drug code or drug name for which the user wishes to receive a notification. When a notification is generated, the notification window is displayed on a monitor containing at least one of the drug name, drug code, content, dosage form, quantity, number of doses, number of days of administration, and reason for notification generation. When multiple drugs are detected simultaneously, the notification order is sorted according to priority criteria, or they are visually displayed differently by classifying them into warning, caution, and reference stages according to importance grades.

[0022] In one embodiment, the system further includes a profile register that generates and stores a user-customized notification drug profile for a drug subject to notification that has been previously selected or entered by a pharmacist, the profile including at least one of a drug name, drug code, ingredient name, content, dosage form, manufacturer information, information on alternative drug groups, warning statement information, and user memo information. The profile register generates the user-customized notification drug profile using a drug name or drug code extracted from an actual display screen of a pharmacy computer system, or generates the user-customized notification drug profile by standardizing a drug name or drug code directly entered by the pharmacist. The judgment and notification unit (700) determines whether a notification is generated by applying at least one multi-stage matching criterion among drug code complete match, drug name similarity match, same ingredient group match, same dosage form group match, and pre-set similar drug group match when comparing the drug information recognized from a newly captured screen with the user-customized notification drug profile, rather than a simple complete match comparison. In particular, if it is determined that a drug corresponding to the user-customized notification drug profile is displayed on the screen, the drug's In addition to the drug name, at least one of the dosage, quantity, number of doses, number of days of administration, reason for warning, user pre-registration memo, and alternative drug guidance information is output together, and the judgment and notification unit (700) accumulates and stores detection history, time information, detection frequency by patient or prescription, and previous notification response status when the same drug or drug of the same ingredient group is repeatedly detected, and is configured to dynamically adjust user-specific notification sensitivity, whether to suppress repeated notifications, priority output drug group or re-emphasize drug group based on the accumulated history information.It can be characterized by going beyond the level of simply reading drugs through OCR to enabling the establishment of a customized monitoring system at the individual pharmacy or individual user level for specific drug groups that pharmacists wish to repeatedly and carefully check in an actual pharmacy work environment.

[0023] The means for solving the problem described above are merely exemplary and should not be interpreted as intended to limit the present invention. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and the detailed description of the invention. Effects of the invention

[0025] According to the present invention, by including a structure that directly receives drug data, including drug name, drug code, ingredient information, and content information, from an external system or generates it from a screen of a pharmacy computer system, it has the effect of being flexibly applicable in both environments where data linkage is possible and environments where it is not. Accordingly, it can be applied without separate structural changes to various pharmacy computer systems, thereby significantly improving system scalability and universality.

[0026] Furthermore, since it is possible to comprehensively analyze the consistency of ingredients, similar drug classes, and contraindications or precautions for concomitant use among multiple drugs based on acquired drug data, it enables precise judgments considering drug interactions beyond simple verification of drug information, thereby having the effect of improving the level of drug safety verification.

[0027] In addition, by determining eligibility for notifications based on pre-set interest drug information or alert conditions by the user, it is possible to provide customized alerts for drugs that pharmacies or users wish to manage as important, thereby improving the work efficiency of pharmacists and supporting decision-making.

[0028] In particular, by dynamically controlling whether to generate notifications or how to output them based on the previous detection history, detection frequency, and user notification response history of the same drug or drugs in the same ingredient group, it is possible to suppress unnecessary repetitive notifications and emphasize necessary notifications, thereby reducing user notification fatigue while improving the efficiency of information delivery.

[0029] In addition, since the reason for the notification is provided along with drug information when outputting notifications, and the output method, frequency, or priority can be variably controlled according to the situation, intuitive and effective information delivery to users is possible, and it is effective to implement an intelligent notification system suitable for actual pharmacy work environments.

[0030] Consequently, the present invention provides, by integrating flexibility in acquiring drug data, precision in analyzing relationships between drugs, user-customized judgment functions, and history-based adaptive notification control functions, the effect of preventing errors in the pharmacy dispensing process, enhancing drug safety, and significantly improving overall work efficiency. Brief explanation of the drawing

[0032] FIG. 1 is a diagram showing the configuration of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 2 is a diagram showing the operation flow of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 3 is a diagram showing the operation sequence of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 4 is a diagram illustrating the area identification, OCR, and structuring processes of a drug data or screen-based relationship analysis and adaptive notification system according to an embodiment of the present invention. FIG. 5 is a diagram showing the operation process of an error control unit of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 6 is a diagram illustrating the drug matching, judgment, and notification output process of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 7 is a diagram illustrating the user profile, multi-stage matching, and dynamic notification control process of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. FIG. 8 is a diagram showing an example of operation of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention. Specific details for implementing the invention

[0033] The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in relation to one embodiment.

[0034] Furthermore, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not intended to be taken in a limiting sense, and the scope of the invention is limited only by the appended claims, including all equivalents thereof, provided appropriately described. Similar reference numerals in the drawings refer to the same or similar functions across various aspects.

[0035] Meanwhile, throughout this specification, when a part is described as “comprising” a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, the “part” for a component as used in this specification performs at least one function or operation. And the “part” may perform the function or operation by hardware, software, or a combination of hardware and software.

[0036] Terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but these components are not limited by the aforementioned terms. The aforementioned terms are used solely for the purpose of distinguishing one component from another.

[0037] In this specification, terms such as “comprising” are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. When a component is referred to as being “connected” to another component, it should be understood that it may be directly connected to or coupled with the other component, or that there may be other components in between.

[0038] Furthermore, in describing the present invention, if it is determined that a detailed description of related known functions or configurations may unnecessarily obscure the essence of the invention, such detailed description is abbreviated or omitted.

[0039] Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings.

[0040] FIG. 1 is a diagram showing the configuration of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0041] FIG. 2 is a diagram showing the operation flow of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0042] FIG. 3 is a diagram showing the operation sequence of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0043] FIG. 4 is a diagram illustrating the area identification, OCR, and structuring processes of a drug data or screen-based relationship analysis and adaptive notification system according to an embodiment of the present invention.

[0044] FIG. 5 is a diagram showing the operation process of an error control unit of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0045] FIG. 6 is a diagram illustrating the drug matching, judgment, and notification output process of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0046] FIG. 7 is a diagram illustrating the user profile, multi-stage matching, and dynamic notification control process of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0047] FIG. 8 is a diagram showing an example of operation of a drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention.

[0048] A drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention comprises: a data acquisition unit (50) configured to receive drug data including a drug name, drug code, ingredient information, and content information, or to generate said drug data from a screen of a pharmacy computer system; a relationship analysis unit (60) that analyzes the relationship between a plurality of drugs based on said drug data; a history-based adaptive risk determination unit (70) that determines whether a drug is a notification target by reflecting the analysis results and user-set conditions; and an adaptive notification output control unit (80) that outputs a notification.

[0049] First, the data acquisition unit (50) may be configured to acquire drug data in various ways. According to one embodiment, the data acquisition unit (50) may receive drug data from an external server, a hospital computer system, or an internal pharmacy database via API calls, file reception, or data linkage. Additionally, according to another embodiment, the data acquisition unit (50) may generate drug data by capturing a screen of a pharmacy computer system, identifying a drug list area from the screen, and performing OCR processing to extract information such as drug name, drug code, content, and quantity. In this way, the data acquisition unit (50) includes both a data reception method and a screen-based generation method, allowing it to be flexibly applied in various system environments.

[0050] The relationship analysis unit (60) is configured to analyze the relationship between multiple drugs based on drug data obtained by the data acquisition unit (50). Specifically, the relationship analysis unit (60) can determine whether the same ingredient is included based on the ingredient information of each drug, and can analyze whether they belong to a similar class based on ingredient groups or pharmacological mechanisms of action. In addition, it can determine whether specific drugs can be used in combination by referring to a drug database in which contraindication information or precaution information for combination use is stored, and can derive risk relationships based on combinations of multiple drugs. This relationship analysis can be extended not only to simple 1:1 comparisons but also to many-to-many analysis of combinations of multiple drugs.

[0051] The above history-based adaptive risk judgment unit (70) is configured to determine whether a notification is required by combining the analysis results of the relationship analysis unit (60) with information on drugs of interest or notification conditions pre-set by the user. At this time, the judgment unit (70) does not simply make a judgment based only on data at the current time, but can dynamically control whether a notification is generated or how a notification is output by considering the previous detection history, detection frequency, and user notification response history of the same drug or drugs of the same ingredient group together.

[0052] Specifically, the judgment unit (70) may determine that a specific drug is a subject for notification if it is included in a pre-set list of drugs of interest or if it corresponds to a contraindication for concomitant use or a subject of caution as a result of relationship analysis. Additionally, if the same drug is detected repeatedly within a short period of time, the system may control the notification to be generated only upon the first detection and subsequently suppress the notification or lower the output intensity. Conversely, if a specific drug is detected repeatedly over a certain period, the system may be configured to increase the importance of the drug to provide a stronger notification. Furthermore, user-customized adaptive control is possible by increasing the priority of a notification when the user repeatedly checks or responds to a specific notification, and conversely, decreasing the frequency of notifications when the user repeatedly ignores them.

[0053] The adaptive notification output control unit (80) is configured to output a notification including drug information and the reason for the notification when a notification is required based on the judgment result of the judgment unit (70). At this time, the notification output control unit (80) is not limited to simple message output, but can dynamically adjust the output method, frequency, and priority of the notification. For example, if the importance is high, a strong form of notification such as a pop-up window, color highlighting, or voice notification can be provided, and if the importance is low, it can be provided in the form of an in-screen display or log record. In addition, if the same notification occurs repeatedly, the notification output frequency can be reduced after a certain number of occurrences or output in a summary form.

[0054] In one embodiment, when the patient’s prescription data includes “Amozaltan 5 / 50mg” and “Vimovojung 5 / 20mg,” the relationship analysis unit (60) determines whether there is an overlap of ingredients between the two drugs or whether there is a caution regarding concomitant use, and if the result matches the user-set condition, the judgment unit (70) can determine it as a notification target. At this time, if there is a history of the drug being repeatedly detected in the prescriptions of the same patient or different patients, the judgment unit (70) can adjust the output method or priority of the notification, and the notification output control unit (80) can provide the user with a specific notification reason, such as “repeated detection of drugs in the same ingredient group” or “detection of drugs requiring caution regarding concomitant use.”

[0055] As such, the system according to the present invention integrates the acquisition of drug data, analysis of relationships between drugs, user-customized judgment, and history-based adaptive notification control, thereby enabling effective management of drug safety without direct linkage with the pharmacy computer system, and providing the effect of improving the work efficiency of pharmacists and preventing errors in the dispensing process.

[0056] A drug data or screen-based relationship analysis and adaptive notification system (10) according to one embodiment of the present invention recognizes drug information displayed on a display screen of a local pharmacy computer without direct API linkage, DB linkage, or internal program modification with the pharmacy computer system, and provides a notification to a pharmacist in real-time or near-real-time based on the notification conditions pre-set by the user or pre-set risk judgment conditions.

[0057] Generally, computer programs used in pharmacies often have a closed structure or limited integration capabilities with external programs, making it difficult for external systems to receive and analyze drug information in real time. In particular, there are limitations in constructing a universal integration module because the types of computer programs, screen configurations, table layouts, drug labeling formats, and code systems used vary from pharmacy to pharmacy. To solve these problems, the present invention provides a configuration that utilizes the screen actually displayed on a monitor by the pharmacy computer system as an information source, extracts drug-related characters from the screen, structures and corrects them, and immediately generates a warning or guidance notification when necessary.

[0058] A system according to one embodiment of the present invention may include a screen capture unit (100), an area identification unit (200), a character recognition unit (300), a data structuring unit (400), an error correction unit (500), a drug matching unit (600), and a judgment and notification unit (700), and may further include an event detection unit (810), a profile registration unit (820), a history storage unit (830), and a user setting unit (840) as needed.

[0059] Each of the above components may be installed and operated on a local pharmacy computer as a single standalone executable program, or may be executed as a background service; additionally, if necessary, some components may operate in conjunction with a separate auxiliary computing device or an internal server. However, the core of the present invention lies in a non-interconnected structure that operates independently from the outside based on screen display results, rather than being directly inserted into or integrated within the pharmacy computer program.

[0060] The present invention relates to a drug data or screen-based relationship analysis and adaptive notification system that recognizes drug information displayed on the screen of a local pharmacy computer without direct data linkage with the pharmacy computer system, analyzes the information, and provides notifications based on conditions pre-set by the user or judgment conditions set within the system. More specifically, the invention relates to a technology that periodically captures the screen of the pharmacy computer system or when a specific event occurs, identifies a region of interest where drug information is displayed on the screen, extracts drug-related information such as drug name, drug code, dosage, dosage quantity, number of doses, and days of administration through Optical Character Recognition (OCR), organizes the extracted information into structured data in row or item units, corrects errors by comparing it with a pre-stored drug database, and outputs a notification if the drug corresponds to a drug of interest to the user or a target for specific risk assessment.

[0061] Generally, computer programs used in pharmacies have closed structures that vary by provider, and in many cases, external applications cannot directly access internal databases or prescription data. Furthermore, some systems do not provide external APIs, or even if they do, their functionality is limited and not standardized, making it difficult to apply them universally to the various computer programs used in the field. In particular, even if a pharmacist wishes to check in real-time during the dispensing process whether a specific drug is included, whether it contains the same active ingredient, whether there are precautions for concomitant use, or whether drugs requiring personal caution are detected, there is a problem in that immediate verification is difficult with existing systems alone, or separate manual work is required.

[0062] To solve such problems, the present invention utilizes the result values ​​actually displayed on the screen as an information source, rather than directly reading internal data from the pharmacy computer system. In other words, since information is acquired based on the screen images output by the computer system on the monitor, it can be applied regardless of the internal structure of a specific computer program, database configuration, integration protocols, or the presence of APIs. Furthermore, by sequentially performing OCR, structuring, correction, and matching based on the drug list displayed on the screen, it is possible to provide drug detection and notification functions at a level suitable for actual use in a pharmacy environment.

[0063] Hereinafter, an embodiment of the present invention will be described in detail, focusing on each component.

[0064] A drug data or screen-based relationship analysis and adaptive notification system according to one embodiment of the present invention may be installed and executed on a local pharmacy computer. Here, the local pharmacy computer may be a general desktop computer, an all-in-one PC, a laptop, or a pharmacy business terminal, and may be the same device on which the pharmacy computer system is executed. In addition, the system of the present invention may be implemented as a separate standalone program, a background resident program, a service-type process, or a combination of an executable file and an auxiliary module.

[0065] The above system functionally includes a screen capture unit (100), an area identification unit (200), a character recognition unit (300), a data structuring unit (400), an error correction unit (500), a drug matching unit (600), and a judgment and notification unit (700). Each of these components may be implemented in a combined form of hardware and software, may be formed as a logic module within a single program, or may be separated into multiple processes or services to operate. For example, screen capture and OCR may be performed on the client side, while drug matching and the application of judgment rules may be performed in a separate internal computation module. However, the essence of the present invention lies in the fact that it is non-interconnected with the pharmacy computer system and that it extracts drug information and provides notifications based on screen display results.

[0066] The above system may operate continuously while the pharmacy computer system displays prescription information, and may be configured to be active only on specific screens as needed. For example, it may operate on a dispensing screen, a prescription confirmation screen, a medication guidance screen, or a pre-payment confirmation screen.

[0067] The screen capture unit (100) is configured to periodically capture the display screen of a pharmacy computer system running on a local pharmacy computer or at the time of a preset event.

[0068] The above screen capture unit (100) can be implemented using an operating system-level screen replication API, a graphic buffer access method, a window handle-based capture method, an active window capture method, a specific monitor capture method, or a method for selecting a coordinate area within the screen. For example, the entire pharmacy computer system window displayed on the front can be acquired as an image frame, or only a specific coordinate area of ​​a pre-designated work screen can be extracted.

[0069] The above screen capture can be performed at regular time intervals. For example, the screen can be checked at a period ranging from 0.1 to 5 seconds, preferably at a period ranging from 0.2 to 1 second. Although the periodic capture method has the advantage of being simple and stable to implement, unnecessary duplicate recognition may occur; therefore, in a preferred embodiment of the present invention, it may be used in parallel with an event-based capture method.

[0070] The pre-configured events here may be, for example, one or more of the following.

[0071] It may be one or more of the following: completion of prescription QR recognition, patient selection or change, completion of prescription calling, creation or updating of a drug list, screen switching, selection of a specific button in the pharmacy computer system, keyboard input or mouse click, change of a specific string or specific color area, change in the number of rows on the screen, or movement of focus on the screen.

[0072] For example, when a pharmacist scans the QR code of a prescription, the prescription list area on the pharmacy computer system is newly updated. At this time, when the event detection function detects the update, the screen capture unit (100) immediately acquires the newly displayed screen. Since this event-based method can perform recognition at the time when the drug information subject to OCR is actually changed, it can ensure both real-time performance and accuracy.

[0073] In addition, the screen capture unit (100) is not limited to capturing only a single frame, but may also continuously acquire preceding and succeeding frames to compare the amount of change. For example, by analyzing the pixel difference between the previous frame and the current frame, changes in character arrangement within a specific area, changes in the number of rows, or changes in the position of the accent color, it is possible to determine whether the drug information has actually changed. This method is useful for preventing repeated OCR on the same screen.

[0074] In one embodiment of the present invention, the pharmacy local computer may be in a multi-display environment with multiple monitors connected. In this case, the screen capture unit (100) may designate only a specific monitor as a capture target or set priorities for each of the multiple monitors. For example, unnecessary recognition can be reduced by targeting only the main monitor displaying the pharmacy computer system and excluding the auxiliary monitor.

[0075] The region identification unit (200) is configured to identify a region of interest in which drug information is displayed from a screen image obtained by the screen capture unit (100) using at least one of coordinate information, screen layout information, color pattern, character arrangement pattern, and arrangement form of user interface objects.

[0076] The screen of a pharmacy computer system generally includes various UI elements in addition to the drug list, such as patient information, insurance information, payment information, button areas, menu bars, and bottom log information. Therefore, if OCR is performed directly on the entire screen, a large amount of unnecessary text is included, which reduces recognition efficiency and increases the possibility of errors. Accordingly, the present invention first identifies the region of interest where drug information is actually placed, and then performs OCR only on that region.

[0077] The area identification unit (200) can primarily perform coordinate-based identification. That is, in the case of a specific pharmacy program, since the drug list area is repeatedly displayed at a fixed screen position, the coordinate values ​​of the area can be set in advance. For example, a rectangular range from the center left to the top right of the screen can be set as the basic area of ​​the drug list.

[0078] Additionally, the area identification unit (200) can perform layout-based identification. For example, in the drug list of a pharmacy program, column headings such as “drug name,” “code,” “specification,” “quantity,” “number of times,” and “days” are often placed in specific positions. In this case, the area identification unit (200) can determine the sub-area as the actual OCR target list based on the detection result of the corresponding column heading string.

[0079] Furthermore, the area identification unit (200) may use color pattern-based identification. Some pharmacy programs highlight the selected row of the drug list with a specific background color such as yellow, blue, or gray, or display the table boundary line with a constant color or line thickness. By utilizing such UI characteristics, the area where the drug list is located can be extracted more reliably.

[0080] Identification based on character array patterns is also possible. For example, if a row structure in which drug names, numbers, and quantities are repeatedly aligned on the same y-axis is detected, it can be determined as a table-format drug list. This method can flexibly respond even when the screen is partially moved or the window size is changed.

[0081] In addition, the area identification unit (200) is not limited to extracting only one large drug list area, but can also internally specify the location of detailed columns or detailed items. For example, the left column can be classified as a drug name column, the right column as a drug code column, the next as a content / specification column, and the right end as a quantity or frequency column. This column-unit identification can be used as basic information for the character recognition unit (300) to apply different recognition conditions corresponding to the character characteristics of each column.

[0082] For example, when multiple rows of drug names are displayed in a table-format area in the center of the screen, and the quantity and number of times are displayed together on the right side of each row, the area identification unit (200) can find the start and end points of the table and calculate the relative interval of each column to individually specify the OCR target section. Through this, non-drug information such as menu text, button text, and total amount can be excluded, and only the actual prescription drug information can be precisely extracted.

[0083] The character recognition unit (300) is configured to generate at least one or more drug-related information, such as a drug name, drug code, dosage, dosage quantity, number of doses, and number of days of administration, by extracting characters, numbers, or symbols included in the area of ​​interest using an optical character recognition method.

[0084] The above character recognition unit (300) may use a public OCR engine, a commercial OCR solution, or a character recognition model separately trained to be suitable for recognizing drug names. Prior to performing OCR, preprocessing work may be performed on the captured image, and such preprocessing may include contrast adjustment, sharpening, binarization, noise removal, tilt correction, enlargement processing, boundary removal, cell separation, etc. Since pharmacy computer screens may contain small text, gray borders, fluorescent highlights, etc. due to monitor rendering characteristics, such preprocessing steps are useful for improving the accuracy of OCR.

[0085] The character recognition unit (300) may perform OCR on the entire region of interest in batches, but preferably, it may recognize parts according to the detailed column information or row information provided by the region identification unit (200). For example, a mixed Korean and English recognition mode may be applied to the drug name region, a number-centered recognition mode may be applied to the drug code region, and a mode that recognizes numbers and unit strings (mg, g, ml, etc.) together may be applied to the content region.

[0086] The drug-related information generated by the character recognition unit (300) may include the following data.

[0087] It may include data such as the drug name, drug code, part or all of the ingredient name, content, specifications, dosage form indicator characters, dosage quantity, frequency of administration, number of days of administration, method of administration, or part of text related to usage.

[0088] For example, if “Amozaltan Tablet 5 / 50mg” and “”””” are each recognized in the first row of the screen, the character recognition unit (300) can output this as one row candidate information. As another example, if “Bimovo Tablet 5 / 20mg” is displayed in the drug name area and “” and “” are displayed in the adjacent number column, the character recognition unit (300) can extract each item individually and transmit it as structured target data.

[0089] Meanwhile, on the pharmacy screen, a single drug name may be displayed as two blocks with line breaks, or the drug name and dosage may be placed adjacently to appear as a single string. In this case, the character recognition unit (300) can generate multiple text blocks containing location information, and the subsequent data structuring unit (400) reorganizes them into a final drug record.

[0090] The data structuring unit (400) is configured to convert drug-related information extracted by the character recognition unit (300) into structured data in rows, columns, or items, and to associate characters located in the same row or adjacent rows with each other to sort into prescription data in individual drug units.

[0091] Since OCR results are generally a sequence of simple strings or a set of text blocks containing positional information, using them as is may result in drug-specific information being processed in a dispersed manner or data from different rows being mixed. Therefore, in this invention, a structuring process that reconstructs the extracted information into actual prescription drug units is crucial.

[0092] The data structuring unit (400) groups blocks belonging to the same row into a single record based on the x-coordinate, y-coordinate, width, height, center point, adjacent distance, and arrangement order of each text block. For example, blocks with the same or similar y-coordinates can be classified as row candidates first, and then aligned from left to right to correspond to each column. At this time, a predetermined tolerance range may be set to account for screen distortion or differences in character height.

[0094] For example, if “Bimoboryeong 5 / 20mg” exists in the left block and “””” exists in the right block in the first row, the data structuring unit (400) can combine them into a single drug record and store it as follows.

[0095] Drug Name: Vimovojung, Content: 5 / 20mg, Quantity: 1, Number of Days or Total: 32

[0096] As another example, if the drug names are recognized across two lines or two blocks as “Tylenol” and “sustained-release tablet,” the data structuring unit (400) can integrate them into a single drug name by considering the positional continuity between the two blocks and the drug name dictionary information.

[0097] Additionally, the data structuring unit (400) can determine what each item means based on column title information or a layout template. For example, if a specific column is learned or set to always correspond to “quantity” and the next column corresponds to “number of times,” each numeric value in the OCR result can be automatically assigned to the corresponding field. Data structured in this way can then be processed more stably in the error correction unit (500) and the drug matching unit (600).

[0098] The error correction unit (500) is configured to correct misrecognition, omission, or separation recognition that occurred during the optical character recognition process by utilizing the correspondence between the drug name, drug code, content, formulation, specification, dosage quantity, number of doses, and number of days of administration included in the structured data, and to improve the accuracy of drug identification by comparing it with standard drug information included in a stored drug database.

[0099] When performing OCR on a pharmacy screen, the following errors may occur.

[0100] Errors such as similar character substitution, missing number digits, missing syllables in drug names, incorrect separation or combination of drug names and content, mixing of data between columns, misrecognition of special characters such as symbols, slashes, and parentheses, and mixing of data from adjacent rows that are not in the same row may occur.

[0101] For example, “Bimoboryeong” may be recognized as “Bimoboryeong,” or “” may be recognized as “”. Additionally, the last digit of the drug code may be omitted, or some syllables in the drug name may be treated as spaces.

[0102] To solve such problems, the error correction unit (500) compares structured data with a stored drug database. The drug database may include standard drug names, drug codes, ingredient names, content, formulations, specifications, manufacturer information, etc., and may also include similar name candidates or information on the same ingredient group as needed.

[0103] The error correction unit (500) can perform correction in the following procedure, for example.

[0104] First is drug code-based correction.

[0105] If drug codes exist in the structured data, code candidates that match or are similar can be searched in the database. Similarity calculations can be performed by considering factors such as missing numbers, number substitutions, and digit errors.

[0106] Second, it is drug name-based correction.

[0107] If a drug name is incomplete or misidentified, standard drug name candidates can be searched using character-level similarity, syllable similarity, edit distance, drug name dictionary matching, etc.

[0108] Third, it is cross-validation-based correction.

[0109] When a drug name, drug code, dosage, and formulation exist simultaneously, the final drug can be identified by comprehensively evaluating the degree of matching among them. For example, even if the drug name is slightly misidentified, if the drug code and dosage exactly match a specific DB item, that DB item can be determined as the final drug.

[0110] Fourth, it is the correction of separation / combination errors.

[0111] If a drug name and dosage are recognized as a single block, or conversely, if a single drug name is recognized as two or more blocks, they can be recombined or reseparated based on location information and DB candidates.

[0112] For example, assume the OCR result is as follows.

[0113] Drug Name: Vimovojang Tablet, Code: 65720570, Strength: 5 / 20mg

[0114] In this case, the error correction unit (500) can confirm in the DB that the standard drug name corresponding to drug code 65720570 is “Bimoborjeong” and, by considering the similarity with the OCR result string, make the final correction to “Bimoborjeong 5 / 20mg”. As such, the present invention can significantly improve recognition accuracy by performing a consistency verification using the mutual correspondence relationship of drug name-code-content-formulation, rather than simply using OCR text.

[0115] The drug matching unit (600) is configured to determine the ingredient information of the drug, whether it has the same ingredient, contraindications for concomitant use, whether it contains specific cautionary ingredients, whether it is a user-designated drug of interest, and whether it is a pre-set notification target by comparing the structured data that has passed through the error correction unit (500) with a previously stored drug database.

[0116] The drug matching unit (600) interprets the meaning of the corresponding drug based on the drug record for which error correction has been completed. It does not simply check the drug name, but connects it with additional information stored in the database to derive the cautionary information or interest information that the pharmacist actually needs.

[0117] For example, the drug matching unit (600) can check the following items.

[0118] Standard ingredient name of the drug, whether duplicate ingredients are included, whether it belongs to a specific drug or ingredient group, whether there are contraindicated concomitant combinations, whether there are combinations requiring caution for concomitant use, whether it is a high-risk drug, whether it is a drug of interest previously registered by the user, whether it matches a specific code or drug name, and whether it exceeds a specific dosage.

[0119] For example, if a pharmacist has set up notifications for a specific product name or a specific ingredient group, the drug matching unit (600) determines whether the identified drug through OCR and correction matches the setting. In addition, even if the product names are different, if they belong to the same ingredient group, they can be determined as drugs of the same family and provided as basic data for subsequent notifications.

[0120] As another example, when multiple drugs detected simultaneously on the screen correspond to a combination of drugs in the DB, the drug matching unit (600) can compare each drug record with one another to generate a matching result in a combination unit.

[0121] The judgment and notification unit (700) determines whether to generate a notification by applying a pre-set risk judgment rule or a user-defined notification rule according to the drug matching result, and if a notification is required, outputs notification information including at least one of a drug name, drug code, dosage, dosage quantity, number of doses, and number of days of administration in the form of a screen, pop-up window, voice, color highlighting, or message.

[0122] The judgment and notification unit (700) first determines whether the condition for generating a notification is met based on the drug matching result. At this time, the notification condition may be a rule built into the system by default, or a user-defined rule set by a user or pharmacy operator.

[0123] For example, a pre-configured judgment rule may be one or more of the following.

[0124] It may be one or more of the following: a notification if a specific drug is included, a notification if a duplicate drug with the same active ingredient is detected, a warning if a contraindicated combination of concomitant use is detected, a caution notification if a specific ingredient or specific dosage is exceeded, a notification if a drug of interest registered by the user is included, or a notification if a drug of a specific risk level or higher is included.

[0125] When the above judgment and notification unit (700) determines that there is a need for notification, it may display a pop-up window in a part of the screen, highlight the row of the corresponding medicine with color, display a warning icon, or output a sound or voice message. Multiple output formats may be used in parallel.

[0126] The notification content may include at least one of the following: drug name, drug code, dosage, quantity, frequency of administration, and days of administration. If necessary, the reason for the warning, ingredient information, user notes, or dispensing precautions may also be displayed.

[0127] For example, if a user has registered “Bimoborb 5 / 20mg” as a drug of interest, when the drug is detected on the screen, the judgment and notification unit (700) can output the following popup.

[0128] Drug of Interest Detected: Vimovojeong 5 / 20mg / Quantity 1 / Days of Administration 32”

[0129] As another example, if two or more drugs with the same active ingredient are detected simultaneously, the following message may be displayed.

[0130] Caution: Requires verification of potential duplication of identical ingredients / Drug A / Drug B”

[0131] Additionally, the judgment and notification unit (700) can assign priority or importance to the notification. For example, high-risk situations such as contraindications for concomitant use can be classified as a warning level, while the detection of drugs of simple interest to the user can be classified as a reference level. Accordingly, the popup color, warning sound, icon, or display location can be varied.

[0132] Below, examples of operation in an actual pharmacy environment are explained in more detail.

[0133] When a pharmacist enters a patient's prescription or recognizes a QR code while the pharmacy computer system is running on the local computer of the pharmacy, multiple drug information is displayed in the central list area of ​​the pharmacy computer screen. At this time, the screen capture unit (100) acquires the corresponding screen. The area identification unit (200) identifies the table-shaped drug list area within the screen, and the character recognition unit (300) extracts text such as drug names, numbers, and dosages from each row.

[0134] For example, “Bimoboryeong 5 / 20mg” in the first row, “” in the second column, and “” in the third column can be extracted. The data structuring unit (400) organizes this into a single drug record, and the error correction unit (500) checks for misrecognition included in the OCR result by comparing it with a standard drug DB. The drug matching unit (600) determines whether the drug is a drug of interest to the user, a subject for judgment of duplicate identical ingredients, or other notification subject based on the corrected result. Finally, the judgment and notification unit (700) outputs a notification window containing the drug name and related information on the monitor when the notification condition is met.

[0135] As such, the present invention can implement the drug notification function necessary during the actual dispensing process without directly accessing internal data of the pharmacy computer program. Furthermore, since it is screen-based, it can be applied to various environments by simply adjusting the screen layout or interest area settings, even if the type of pharmacy program changes.

[0136] According to the configuration of the present invention, first, since drug information displayed on the screen can be utilized without direct linkage with the pharmacy computer system, it has the advantage of being applicable to existing programs with a closed structure. Second, by performing structuring and error correction beyond simple OCR, it can secure a level of reliability applicable to actual pharmacy operations. Third, through matching with a drug database and applying judgment rules, it can provide practical operational support functions beyond simple drug detection, such as drugs of caution, drugs of interest, identical ingredients, and concomitant risks. Fourth, since the notification content can display not only the drug name but also the dosage, quantity, frequency of administration, and days of administration, pharmacists can immediately check necessary information and reflect it in their dispensing decisions.

[0137] Furthermore, the configuration of the present invention can be extended in the future to include user-customized profile registration, management of repeated detection history, learning of drug-specific priorities, automatic expansion of similar drugs, enhanced warnings to prevent dispensing errors, and identification of drugs subject to medication guidance. However, despite such expansion, the core of the present invention lies in a non-interconnected system structure that directly recognizes the pharmacy computer system screen to structure and correct drug information and provides information to the user according to notification conditions.

[0139] In an embodiment of the present invention, the screen capture unit (100) is not limited to simply capturing the screen at regular time intervals, but is configured to detect an event directly linked to a change in prescription information of the pharmacy computer system and perform the capture.

[0140] The above event is caused by a change in the internal state of the pharmacy computer system or user input, and may include, for example, the following cases.

[0141] This may include the time when prescription QR code recognition is completed, the time when patient selection or change is made, the time when prescription data retrieval is completed, the time when the drug list is updated, the time when switching to the dispensing screen or medication guidance screen, the time when data is finalized following a specific shortcut key input (e.g., prescription confirmation key), a mouse click, or an Enter key input, a change in the pixel value of text or a specific area within the screen, or a change in the number of rows in the drug list.

[0142] In this case, event detection can be implemented through operating system-level input event detection, screen comparison methods, UI object change detection methods, or image hash comparison methods for specific areas.

[0143] According to one embodiment of the present invention, the screen capture unit (100) may be configured to compare the screen before and after the occurrence of an event and to perform OCR processing only when data within the drug list area has actually changed. For example, subsequent processing may be performed only when a change in a string containing a drug name, an increase or decrease in the number of rows, or a change in a specific cell value is detected by comparing a specific area of ​​the previous frame and the current frame.

[0144] This prevents the repetitive execution of OCR on the same screen, minimizes system resource usage, and reduces duplicate notifications. Furthermore, by performing recognition at the actual time of prescription change, both real-time performance and accuracy can be ensured simultaneously.

[0145] For example, OCR is not performed repeatedly while the pharmacist maintains the same patient's screen and performs other tasks, and recognition is performed only when a new prescription is called or the drug list is changed.

[0147] In an embodiment of the present invention, the area identification unit (200) more precisely identifies the area where drug information is displayed within the pharmacy computer system screen, and the character recognition unit (300) and the data structuring unit (400) perform recognition and structuring according to the characteristics of the area.

[0148] The above area identification unit (200) can distinguish the following detailed areas within the screen.

[0149] You can distinguish the entire drug list area, drug name column area, drug code column area, content or specification column area, quantity column area, dosage frequency column area, dosage days column area, and column title area.

[0150] The above area division can be performed based on coordinates, layout, character patterns, or color patterns, and one or more methods may be used in combination.

[0151] The character recognition unit (300) may apply different OCR recognition conditions according to the characteristics of each area. For example, a mixed Korean and English recognition mode may be applied to the drug name area, a number-centered recognition mode may be applied to the drug code area, a mixed recognition mode of numbers and unit strings (mg, g, ml, etc.) may be applied to the content area, and a number-first recognition mode may be applied to the quantity, frequency, and days areas.

[0152] Through this, recognition results optimized for the character characteristics of each area can be obtained.

[0153] The data structuring unit (400) sorts text blocks extracted from the OCR result according to location information and combines texts belonging to the same row into a single drug data record. At this time, the determination of the same row is performed based on the y-coordinate or center point coordinate of the text block, and blocks within a certain error range may be considered as the same row.

[0154] Additionally, the data structuring unit (400) can map each item to a field based on the relative position of each column. For example, the first column from the left can be set as the drug name, the second column as the code, the third column as the content, the fourth column as the quantity, etc.

[0155] For example, assume that the following OCR results exist.

[0156] (x1, y1): “Amozaltan Tablet 5 / 50mg”, (x2, y1): “”, (x3, y1): “”, (x4, y1): “”

[0157] The data structuring unit (400) can group this into one record and structure it as: drug name: Amozaltan tablet, content: 5 / 50 mg, quantity: 1, number of doses: 2, number of days of dose: 30.

[0158] In addition, if a drug name is separated into two or more blocks, it can be combined into a single drug name by considering the distance between adjacent blocks and string patterns.

[0160] In an embodiment of the present invention, the error correction unit (500) ensures consistency by comparing multiple pieces of information, including drug codes and drug names, to correct various recognition errors that occur during the OCR process.

[0161] The following errors can frequently occur during the OCR process.

[0162] Confusion between numbers and letters, confusion between numbers and letters, misrecognition of units such as “mg” and “”, omission or alteration of parts of Korean syllables, occurrence of spaces in drug names, errors in combining or separating drug names and dosages

[0163] The error correction unit (500) compares the structured data with the stored drug database to correct these errors.

[0164] The above correction process can be performed in the following steps.

[0165] First, drug code-based candidate search: deriving candidates by calculating similarity between codes extracted by OCR and codes in the database, and considering digit errors, numerical substitutions, etc.

[0166] Second, drug name-based candidate search: deriving candidates based on string similarity, edit distance, and syllable similarity, utilizing drug name dictionaries or stored lists.

[0167] Third, cross-validation: evaluation of consistency between drug name, code, strength, and formulation, and selection of the candidate with the highest consistency.

[0168] For example, assume the OCR result is as follows.

[0169] Drug Name: “Bimovojang Tablet”, Code: “”, Strength: “”

[0170] In this case, the error correction unit (500) searches the DB to confirm that the standard drug name corresponding to the code “ ” is “Bimoborjeong”, and compares the similarity with the OCR result to finally correct it to “Bimoborjeong 5 / 20mg”.

[0171] In addition, when referring to drugs with different names and codes, the final result can be determined based on code priority, name priority, or user-defined criteria.

[0172] This correction process substantially improves OCR accuracy and plays an important role in ensuring the reliability of subsequent drug matching and notification judgments.

[0174] In an embodiment of the present invention, the drug matching unit (600) and the judgment and notification unit (700) determine whether there is a risk or interest based on various conditions according to the drug database and user setting information, and provide the result to the user.

[0175] The above judgment criteria may include one or more of the following.

[0176] Whether there is duplicate prescription of drugs with the same active ingredient, whether specific ingredients are included, whether there are combinations with contraindications or precautions for concomitant use, whether there is duplicate inclusion of drugs in the same class, whether it is a drug of caution for specific age or disease groups, whether the established standard dosage has been exceeded, whether it is a drug of interest registered by the user, and whether it matches a specific drug code or drug name.

[0177] For example, if two or more drugs having the same ingredients are detected simultaneously, the judgment and notification unit (700) may determine this as a possibility of duplicate prescription and output a warning notification.

[0178] In addition, if a user has registered a specific drug as a drug of interest, an alert can be generated immediately when that drug is detected.

[0179] Notifications may be displayed in the following forms: pop-up windows, on-screen highlighting (color change, border display, etc.), warning icon display, sound or voice notifications, message windows

[0180] The notification content may include the following information: drug name, drug code, dosage, quantity, frequency of administration, days of administration, and reason for the notification.

[0181] For example, if a specific drug is registered as a drug of interest to the user, the following notification may be displayed.

[0182] Drug of Interest Detected: Vimovojung 5 / 20mg / Quantity 1 / 32 Tablets”

[0183] In addition, if a combination requiring caution is detected, a message such as “Caution: Drug combination requiring caution detected / Drug A + Drug B” may be displayed.

[0184] In addition, when multiple notifications occur simultaneously, priorities can be set according to importance to display them in warning, caution, and reference stages.

[0185] With this configuration, real-time performance can be ensured while reducing unnecessary computations through event-based capture; flexibility can be achieved in various pharmacy program environments through UI area identification and structuring; accuracy suitable for actual work applications can be secured through OCR error correction; and information practically needed by pharmacists can be provided immediately through various condition-based judgment logics.

[0186] Consequently, the present invention provides technical effects that can contribute to verifying drug safety, preventing dispensing errors, and improving work efficiency, while overcoming the structural limitations of pharmacy computer systems.

[0187] In one embodiment of the present invention, the system may further include a profile register that generates user-customized notification drug profiles and performs notifications based thereon, in order to provide a customized notification function that can be set at the user or pharmacy level, going beyond general drug matching and notification functions based on a stored drug database. The profile register is configured to allow a pharmacist to register specific drugs or drug groups that they wish to carefully check during the actual dispensing process, and the registered profile may include at least one piece of information such as drug name, drug code, ingredient name, dosage, formulation, manufacturer information, information on the same ingredient group or similar drug group, information on alternative drug groups, user memo, notification importance, notification method, and whether repeated notifications are allowed.

[0188] The aforementioned user-customized notification drug profile can be created by a pharmacist directly entering the drug name or drug code, or it can be automatically generated based on drug information recognized via OCR from the pharmacy computer system screen. For example, if a pharmacist identifies a specific drug during the dispensing process and wishes to receive notifications regarding it in the future, they can select the drug through the system interface to register it as a notification target; in this case, the drug name and code extracted via OCR can be automatically reflected in the profile. Consequently, users can naturally configure notification profiles during their actual workflow without the burden of separate input.

[0189] The above judgment and notification unit (700) can determine whether to generate a notification by applying multiple matching criteria when comparing the newly recognized drug information through screen capture and OCR processes with the user-customized notification drug profile, rather than judging based solely on whether there is a complete match of a simple drug name or drug code. Specifically, in the first step, a judgment is made based on a complete match of the drug code or a complete match of the drug name; in the second step, similar matching is performed considering the similarity of the drug name string, edit distance, or syllable similarity; in the third step, whether the same ingredient is included or whether there is a same ingredient group is determined based on the drug ingredient information; in the fourth step, whether there is a same formulation or a similar dosage range is determined based on the formulation and content information; and in the fifth step, matching can be performed for a similar drug group, an alternative drug group, or a specific condition-based group predefined by the user.

[0190] Through this multi-stage matching structure, notifications can be provided that include not only the same product but also other products with the same ingredients or drugs of similar classes, enabling the more effective detection of risk factors that are difficult to identify by product name alone in an actual pharmacy environment.

[0191] In addition, the judgment and notification unit (700) may be configured to not only simply indicate the presence or absence of the drug when a notification occurs, but also to output various information included in the user-customized notification drug profile. For example, when a notification is output, not only basic information such as the drug name, drug code, dosage, quantity, number of doses, and number of days of administration, but also memo information pre-entered by the user, reason for warning, information on alternative drugs, or information on the same ingredient group may be displayed together. Accordingly, the pharmacist can immediately check additional judgment information regarding the drug at the time of notification, and can make faster and more accurate decisions during the medication guidance or dispensing judgment process.

[0192] Furthermore, according to one embodiment of the present invention, the system may include a history storage function that stores a history of repeated detections of the same drug or drugs of the same ingredient group and dynamically controls the notification method based thereon. The history storage function accumulates and stores data including the time of detection, the name of the detected drug, the drug code, the number of detections, patient or prescription identification information, whether a notification occurred, and whether a user responded. Based on the stored history data, the judgment and notification unit (700) can dynamically adjust the output method, frequency, and priority of the notification.

[0193] For example, if the same drug is detected repeatedly within a short period, unnecessary repetitive notifications can be suppressed by displaying a pop-up notification only once and maintaining only a screen highlight thereafter; conversely, if a specific drug is detected repeatedly over a certain period, that drug can be determined as an important drug, thereby increasing the priority of the notification or providing a stronger visual or auditory alert. Additionally, user-customized notification sensitivity can be adjusted by reducing the frequency or importance of a specific notification if the user repeatedly ignores it, while increasing its priority if the user continuously checks or responds to it.

[0194] For example, when a pharmacist has registered an alert profile for a specific drug that includes a note stating “needs enhanced medication guidance,” and that drug is displayed on the prescription screen, the system can output an alert in the form of “Drug of interest detected: Bimovojung 5 / 20mg, Quantity 1, Days of administration 32, User note: Needs enhanced medication guidance.” Subsequently, if the same drug is detected repeatedly within the same prescription, the system maintains only the screen highlighting without generating an additional popup; conversely, if the same drug is detected repeatedly in prescriptions for various patients, it can be classified as a repeatedly detected drug, thereby increasing the importance of the alert.

[0195] As described above, according to the configuration of the present invention, by combining a user-customized profile, a multi-stage matching structure, and history-based learning-type notification control beyond a simple OCR-based drug recognition system, it is possible to implement an intelligent drug monitoring and notification system optimized for actual pharmacy work environments, which can significantly contribute to ensuring drug safety, preventing dispensing errors, and improving work efficiency.

[0196] Although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above. Various modifications are possible by those skilled in the art without departing from the essence of the invention as claimed in the claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention. Explanation of the symbols

[0198] 10. Drug data or screen-based relationship analysis and adaptive notification system

Claims

Claim 1 A data acquisition unit (50) configured to receive drug data including drug name, drug code, ingredient information, and content information, or to generate said drug data from a screen of a pharmacy computer system; a relationship analysis unit (60) that analyzes relationships between drugs based on said drug data, including whether ingredients match, whether they belong to similar classes, and relationships of contraindications or precautions for concomitant use between multiple drugs; and a history-based adaptive risk judgment unit (70) that determines whether to be a notification target by combining the results of said relationship analysis with drug information of interest or notification conditions pre-set by the user, and dynamically controls whether to generate a notification or the method of outputting a notification by reflecting the previous detection history, detection frequency, and user notification response history of the same drug or drugs of the same ingredient group. ...and if notification is required according to the above judgment result, an adaptive notification output control unit (80) outputs a notification including drug information and a reason for notification, and dynamically adjusts the output method, frequency, or priority of the notification; a screen capture unit (100) that periodically captures a display screen of a pharmacy computer system running on a local computer of a pharmacy or at the time of a preset event occurrence; a region identification unit (200) that identifies a region of interest where drug information is displayed from a screen image obtained by the screen capture unit (100) using at least one of coordinate information, screen layout information, color pattern, character array pattern, and arrangement form of user interface objects; a character recognition unit (300) that extracts characters, numbers, or symbols included in the region of interest using an optical character recognition method to generate at least one or more drug-related information such as drug name, drug code, content, dosage quantity, number of doses, and number of days of administration; and data that converts the drug-related information extracted by the character recognition unit (300) into structured data in row units, column units, or item units, and associates characters located in the same row or adjacent rows to sort them into prescription data for individual drugs. Structured section (400);Error correction unit (500) that corrects misrecognition, omission, or separation recognition occurring during the optical character recognition process by utilizing the correspondence between the drug name, drug code, content, formulation, specification, dosage quantity, number of doses, and number of days of administration included in the above structured data, and improves the accuracy of drug identification by comparing it with standard drug information included in a stored drug database; drug matching unit (600) that determines the ingredient information, whether the ingredients are identical, contraindications for concomitant use, whether specific caution ingredients are included, whether it is a user-designated drug of interest, and whether it is a pre-set notification target by comparing the structured data that has passed through the above error correction unit (500) with the stored drug database; The system further includes a judgment and notification unit (700) that determines whether to generate a notification by applying a preset risk judgment rule or a user-defined notification rule according to the above drug matching result, and outputs notification information including at least one of a drug name, drug code, dosage, dosage quantity, number of doses, and number of days of administration in the form of a screen, pop-up window, voice, color highlighting, or message if a notification is required; wherein the screen capture unit (100) targets at least one of a prescription input screen, dispensing screen, medication guidance screen, payment screen, or drug list display screen of a pharmacy computer system to capture; wherein the preset event includes at least one of completion of prescription QR recognition, prescription data retrieval, screen switching, cursor position change, drug list update, change of dispensing patient, input of a specific shortcut key, mouse click, change of string within the screen, change of a specific color area, or activation of a specific UI object; and wherein the screen capture unit (100) is configured to perform a subsequent recognition procedure only when it is confirmed that the drug list has changed by comparing the screen before and after the occurrence of the event, characterized in that it is a drug data or screen-based relationship analysis and adaptive notification. System.; Claim 2 delete Claim 3 delete Claim 4 In claim 1, the area identification unit (200) distinguishes a table area, list area, cell area, column title area, quantity display area, and total display area where drug information is displayed within the screen; the character recognition unit (300) applies different character recognition conditions to each identified area, applying a Korean priority recognition condition to the Korean drug name area, a number priority recognition condition to the drug code area, and a mixed recognition condition of numbers and unit strings to the content and specification area, respectively; the data structuring unit (400) groups adjacent text blocks on the same horizontal line or within a preset error range into a single drug row, stores the drug name, drug code, content, quantity, number of doses, and number of days of dose extracted for each row by corresponding them to each attribute field, and is configured to individually generate a structured record for each drug when multiple drugs are listed, characterized in that the drug data or screen-based relationship analysis and adaptive notification system. Claim 5 In paragraph 4, the error correction unit (500) performs a cross-comparison based on standard drug name, drug code, ingredient name, dosage form, content, and specification information included in a stored drug database when either the drug name or drug code extracted as a result of optical character recognition is incomplete or misrecognized; when the drug code is recognized with only some digits or replaced with similar numbers, it derives a correction candidate using a number sequence similarity and a preset digit pattern; when the drug name is separated into some syllable units or misrecognized with similar characters, it derives a correction candidate using string similarity, a stored drug dictionary, preset prohibited words, and a drug name prefix or suffix pattern; among the correction candidates, it confirms a candidate whose degree of agreement with the drug name, drug code, content, and dosage form is greater than or equal to a reference value as the final drug information; and when the drug name and drug code indicate different drugs, it sets the drug code as the priority standard or selectively applies one of the drug name priority standard, code priority standard, or multiple candidate warning output standard according to user settings. A drug data or screen-based relationship analysis and adaptive notification system characterized by being configured. Claim 6 In claim 5, the drug matching unit (600) and the judgment and notification unit (700) determine whether to generate a notification based on at least one of the following criteria, using a drug included in a stored drug database or a user-registered notification list: whether there is a duplicate prescription of the same ingredient, whether there is a pre-set specific ingredient, whether there is a combination of contraindications or precautions for concomitant use, whether there is a duplicate of drugs of the same class, whether there is a drug of caution for a specific age group or specific disease group, whether there is an excess of a pre-set dosage or pre-set quantity, whether there is a drug of interest registered in advance by the user, and whether there is a match with the drug code or drug name for which the user wishes to receive a notification; when a notification is generated, a notification window is displayed on a monitor including at least one of the drug name, drug code, content, dosage form, quantity, number of doses, number of days of administration, and reason for notification generation of the corresponding drug; and when multiple drugs are detected simultaneously, the notification order is sorted according to priority criteria or is visually displayed differently by classifying them into warning, caution, and reference stages according to importance grades. This characterizes the drug data or screen-based relationship analysis and adaptive notification. System. Claim 7 In claim 6, the system further includes a profile register that generates and stores a user-customized notification drug profile for a drug subject to notification selected or entered by a pharmacist in advance, the profile register including at least one of a drug name, drug code, ingredient name, content, dosage form, manufacturer information, information on alternative drug groups, warning statement information, and user memo information. The profile register generates the user-customized notification drug profile using a drug name or drug code extracted from an actual display screen of the pharmacy computer system, or generates the user-customized notification drug profile by standardizing and processing a drug name or drug code directly entered by the pharmacist. The judgment and notification unit (700) determines whether a notification occurs by applying at least one multi-stage matching criterion among a drug code complete match, drug name similarity match, same ingredient group match, same dosage form group match, and pre-set similar drug group match when comparing the drug information recognized from a newly captured screen with the user-customized notification drug profile, rather than a simple complete match comparison. In particular, when it is determined that a drug corresponding to the user-customized notification drug profile is displayed on the screen, The system is characterized by outputting not only the name of the drug but also at least one of the dosage, quantity, number of doses, number of days of administration, reason for warning, user pre-registration memo, and alternative drug guidance information together, and the judgment and notification unit (700) accumulates and stores detection history, time information, detection frequency by patient or prescription, and previous notification response status when the same drug or drug of the same ingredient group is repeatedly detected, and is configured to dynamically adjust user-specific notification sensitivity, whether to suppress repeated notifications, priority output drug group or re-emphasize drug group based on the accumulated history information, thereby going beyond the level of simply reading drugs through OCR to enable the establishment of a customized monitoring system at the individual pharmacy or individual user level for specific drug groups that pharmacists want to repeatedly and carefully check in an actual pharmacy work environment.Drug data or screen-based relationship analysis and adaptive notification system.

Images