Automatic data acquisition method and device

Abstract

The invention discloses an automatic collection method and device, relates to the technical field of medical informatization, and can solve the problems of low efficiency and high error probability when blood glucose data is collected in a manual recording mode in a hospital at the present stage. Inputting the to-be-collected data into a hospital information system (HIS) according to the automatic input script; in the process of inputting the to-be-collected data into the HIS, according to the automatic task step recording script, the input control function of the to-be-collected data is achieved. The method is used for achieving efficient and accurate input of patient data in the medical field, and the automation level and reliability of data collection are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of medical information technology, and in particular to a method and apparatus for automatic data acquisition. Background Technology

[0002] Optical Character Recognition (OCR) is a technology that converts text (including printed and handwritten text) in images into editable and searchable digital text. This technology typically uses electronic devices such as scanners and digital cameras to detect characters in digital images or paper documents, and then uses pattern recognition technology to convert the detected characters into computer-editable text.

[0003] In the medical field, traditional blood glucose data collection generally includes two methods: manual entry and Bluetooth transmission. Manual entry typically involves using a standard blood glucose meter to measure the patient's blood glucose level. Medical staff then manually record the measured values ​​on paper forms and manually input them into the Hospital Information System (HIS), a rather cumbersome process. While Bluetooth transmission eliminates the need for manual recording of patient blood glucose values, it requires a specialized Bluetooth-enabled blood glucose meter and a customized interface with the hospital information system for convenient data transmission, making it more expensive.

[0004] Currently, due to limitations in medical infrastructure and usage habits, most hospitals still rely on manual data entry when collecting blood glucose data. This method is inefficient and prone to errors due to human error. Summary of the Invention

[0005] This application provides an automatic data acquisition method and apparatus, which can solve the problems of low efficiency and high error rate when hospitals collect blood glucose data manually.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] In a first aspect, this application provides an automatic data collection method, comprising: acquiring data to be collected; entering the data to be collected into a hospital information system (HIS) according to an automated data entry script; and recording a script according to the automated task steps during the process of entering the data to be collected into the HIS, thereby realizing the data entry control function.

[0008] In one possible implementation, the data to be collected is entered into the HIS according to the automated data entry script, specifically including: simulating HIS operation instructions according to the automated data entry script; and entering the data to be collected into the HIS according to the HIS operation instructions.

[0009] In one possible implementation, the data entry control function includes at least one of the following sub-functions: step recording function, step pause function, step continue function, specified step start function, log viewing function, and blood glucose alarm function; wherein, the blood glucose alarm function is used to generate and display alarm information when there is an abnormal blood glucose value in the data to be collected.

[0010] In one possible implementation, the type of data to be collected is blood glucose data. Obtaining the data to be collected specifically includes: obtaining a blood glucose record sheet; wherein the data category recorded on the blood glucose record sheet includes at least one of the following: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value; and obtaining the data to be collected from the blood glucose record sheet through optical character recognition.

[0011] In one possible implementation, data to be collected is obtained from a blood glucose record based on optical character recognition, specifically including: taking a picture of the blood glucose record to obtain image data of the blood glucose record; determining the data to be collected from the image data of the blood glucose record based on a preset algorithm; wherein the preset algorithm includes at least one of the following: convolutional neural network (CNN), attention mechanism, and character segmentation algorithm.

[0012] In one possible implementation, when the preset algorithm includes a convolutional neural network, an attention mechanism, and a character segmentation algorithm, the data to be collected is determined from the image data of the blood glucose record based on the preset algorithm. Specifically, this includes: extracting features from the image data of the blood glucose record using a convolutional neural network to obtain character feature data; determining the target character based on the attention mechanism and the character feature data, and semantically associating the target character; and performing single-character separation on the semantically associative target character using the character segmentation algorithm to obtain the data to be collected.

[0013] In one possible implementation, feature extraction is performed on the image data of the blood glucose record using a convolutional neural network to obtain character feature data, which is then executed using the following formulas one and two:

[0014]

[0015] in, This represents the value of the feature map output by the l-th convolutional layer at position (i,j). This represents the weight of the l-th convolutional kernel at position (m, n). b represents the value of the input image or the feature map of the previous layer at (i+m,j+n). (l) k represents the bias term of the l-th layer. h k represents the height of the convolution kernel. w Indicates the width of the convolution kernel;

[0016]

[0017] Where, p h p represents the height of the pooling window. w The width of the pooling window is represented by s, and the step size is represented by s.

[0018] Based on the attention mechanism and character feature data, the target character is determined, and semantic association is performed on the target character using the following formulas three and four:

[0019]

[0020] Where, α i,j Q represents the attention weight for the i-th query corresponding to the j-th key. i K represents the query vector. i This represents the key vector, and score represents the scoring function;

[0021]

[0022] Among them, C i V represents the context vector at position i. j Represents a value vector;

[0023] Based on the character segmentation algorithm, the target characters after semantic association are separated into individual characters to obtain the data to be collected, which is then executed using the following formulas five and six:

[0024]

[0025] Where I(x,y) represents the pixel value of the binarized image at coordinates (x,y), H(x) represents the vertical projection histogram, V(y) represents the horizontal projection histogram, w represents the width of the image, and h represents the height of the image;

[0026]

[0027] Where L(x,y) represents the connected component label of pixel (x,y), new_label represents the newly discovered connected component number, and existing_label represents the existing label of adjacent pixels.

[0028] In one possible implementation, before entering the data to be collected into the HIS according to the automated data entry script, the method further includes: generating an automated data entry script and an automated task step recording script; and performing passability tests on the automated data entry script and the automated task step recording script.

[0029] Secondly, this application provides an automatic data acquisition system, applied to the automatic data acquisition method described in any possible implementation of the first aspect. The automatic data acquisition system includes: a data acquisition module, an automated data entry module, a data entry control module, and a script generation and testing module. The data acquisition module is used to acquire data to be acquired. The automated data entry module is used to enter the data to be acquired into the Hospital Information System (HIS) according to an automated data entry script. The data entry control module is used to record scripts according to automated task steps during the process of entering the data to be acquired into the HIS, thereby implementing the data entry control function. The script generation and testing module is used to generate automated data entry scripts and automated task step recording scripts, and to perform pass / fail tests on the automated data entry scripts and automated task step recording scripts.

[0030] Thirdly, this application provides an automatic data acquisition device, comprising: an acquisition unit and a processing unit; the acquisition unit is used to acquire data to be acquired; the processing unit is used to input the data to be acquired into the Hospital Information System (HIS) according to an automated input script; the processing unit is also used to record a script according to the automated task steps during the process of inputting the data to be acquired into the HIS, thereby realizing the data input control function.

[0031] In one possible implementation, the processing unit is further configured to simulate HIS operation instructions according to the automated data entry script; the processing unit is further configured to enter the data to be collected into the HIS according to the HIS operation instructions.

[0032] In one possible implementation, the data entry control function includes at least one of the following sub-functions: step recording function, step pause function, step continue function, specified step start function, log viewing function, and blood glucose alarm function; the processing unit is also used to generate and display alarm information when there are abnormal blood glucose values ​​in the data to be collected.

[0033] In one possible implementation, the acquisition unit is further configured to acquire a blood glucose record sheet; wherein the data categories recorded on the blood glucose record sheet include at least one of the following: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value; the processing unit is further configured to acquire the data to be collected from the blood glucose record sheet through optical character recognition.

[0034] In one possible implementation, the acquisition unit is further configured to capture an image of the blood glucose record sheet to acquire image data of the blood glucose record sheet; the processing unit is further configured to determine the data to be collected from the image data of the blood glucose record sheet based on a preset algorithm; wherein the preset algorithm includes at least one of the following: convolutional neural network (CNN), attention mechanism, and character segmentation algorithm.

[0035] In one possible implementation, the processing unit is further configured to extract features from the image data of the blood glucose record sheet using a convolutional neural network to obtain character feature data; the processing unit is further configured to determine the target character based on the attention mechanism and the character feature data, and to perform semantic association on the target character; the processing unit is further configured to perform single-character separation on the target character after semantic association using a character segmentation algorithm to obtain the data to be collected.

[0036] In one possible implementation, the processing unit extracts features from the image data of the blood glucose record using a convolutional neural network to obtain character feature data, which is then executed using the following formulas one and two:

[0037]

[0038] in, This represents the value of the feature map output by the l-th convolutional layer at position (i,j). This represents the weight of the l-th convolutional kernel at position (m, n). b represents the value of the input image or the feature map of the previous layer at (i+m,j+n). (l) k represents the bias term of the l-th layer. h k represents the height of the convolution kernel. w Indicates the width of the convolution kernel;

[0039]

[0040] Where, p h p represents the height of the pooling window. w The width of the pooling window is represented by s, and the step size is represented by s.

[0041] The processing unit determines the target character based on the attention mechanism and character feature data, and performs semantic association on the target character, which is executed through the following formulas three and four:

[0042]

[0043] Where, α i,j Q represents the attention weight for the i-th query corresponding to the j-th key. i K represents the query vector. i This represents the key vector, and score represents the scoring function;

[0044]

[0045] Among them, C i V represents the context vector at position i. j Represents a value vector;

[0046] The processing unit performs single-character separation on the target characters after semantic association according to the character segmentation algorithm to obtain the data to be collected, which is then executed using the following formulas five and six:

[0047]

[0048] Where I(x,y) represents the pixel value of the binarized image at coordinates (x,y), H(x) represents the vertical projection histogram, V(y) represents the horizontal projection histogram, w represents the width of the image, and h represents the height of the image;

[0049]

[0050] Where L(x,y) represents the connected component label of pixel (x,y), new_label represents the newly discovered connected component number, and existing_label represents the existing label of adjacent pixels.

[0051] In one possible implementation, the processing unit is further configured to generate automated data entry scripts and automated task step recording scripts; the processing unit is further configured to perform passability tests on the automated data entry scripts and automated task step recording scripts.

[0052] Fourthly, this application provides a computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by an electronic device of this application, cause the electronic device to perform the automatic data acquisition method as described in the first aspect and any possible implementation thereof.

[0053] Fifthly, this application provides an electronic device, including: a processor and a memory; wherein the memory is used to store one or more programs, the one or more programs including computer-executable instructions, and when the electronic device is running, the processor executes the computer-executable instructions stored in the memory to cause the electronic device to perform the automatic data acquisition method as described in the first aspect and any possible implementation thereof.

[0054] Sixthly, this application provides a computer program product containing instructions that, when executed on a computer, cause the electronic device of this application to perform the automatic data acquisition method as described in the first aspect and any possible implementation thereof.

[0055] In a seventh aspect, this application provides a chip system applied to an automatic data acquisition device; the chip system includes one or more interface circuits and one or more processors. The interface circuits and the processors are interconnected via lines; the interface circuits are used to receive signals from the memory of the automatic data acquisition device and send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the automatic data acquisition device performs the automatic data acquisition method as described in the first aspect and any possible design of the first aspect.

[0056] Based on the above technical solution, this application combines optical character recognition technology with automated scripts to automatically identify key data on paper patient data records, avoiding manual entry errors and reducing data deviations. Utilizing automated scripts to simulate keyboard and mouse operations, it can accurately input patient data into the HIS system. Furthermore, because the input time for a single data entry is reduced from minutes to seconds, this application's technical solution is suitable for batch data entry scenarios. Simultaneously, this application also supports functions such as step pause, specified step start, and data anomaly alarms, adapting to the manual review needs in medical scenarios. Therefore, this application's technical solution achieves efficient and accurate input of patient data in the medical field without requiring additional hardware, significantly improving the automation level and reliability of data collection. Attached Figure Description

[0057] Figure 1 This application provides a schematic diagram of the architecture of an automatic data acquisition system.

[0058] Figure 2 A flowchart illustrating an automatic data acquisition method provided in an embodiment of this application;

[0059] Figure 3 A flowchart illustrating another automatic data acquisition method provided in an embodiment of this application;

[0060] Figure 4 A flowchart illustrating another automatic data acquisition method provided in an embodiment of this application;

[0061] Figure 5 A flowchart illustrating another automatic data acquisition method provided in an embodiment of this application;

[0062] Figure 6 This is a schematic diagram of the structure of an automatic data acquisition device provided in an embodiment of this application;

[0063] Figure 7 This is a schematic diagram of another automatic data acquisition device provided in an embodiment of this application. Detailed Implementation

[0064] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0065] In this article, the character " / " generally indicates that the objects before and after it are in an "or" relationship. For example, A / B can be understood as A or B.

[0066] The terms "first" and "second" in the specification and claims of this application are used to distinguish different objects, not to describe a specific order of objects. For example, "first edge service node" and "second edge service node" are used to distinguish different edge service nodes, not to describe a characteristic order of edge service nodes.

[0067] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0068] Furthermore, in the embodiments of this application, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplarily" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplarily" or "for example" is intended to present concepts in a concrete manner.

[0069] The following is a description of the technical terms used in this application:

[0070] 1. Optical Character Recognition

[0071] Optical Character Recognition (OCR) is a technology that converts text (including printed and handwritten text) in images into editable and searchable digital text. This technology typically uses electronic devices such as scanners and digital cameras to detect characters in digital images or paper documents, and then uses pattern recognition technology to convert the detected characters into computer-editable text.

[0072] Generally speaking, OCR technology includes two main steps:

[0073] Step 1: Use electronic devices such as scanners and digital cameras to take pictures of paper documents or digital images, and perform image preprocessing on the captured images. The preprocessing process may specifically include operations such as denoising, binarization, and layout analysis.

[0074] Step 2: Perform text detection and recognition on the preprocessed images. At present, when it comes to text detection and recognition, the mainstream solution is to implement it through deep learning methods, such as end-to-end models (extracting features through convolutional neural networks, performing sequence modeling through recurrent neural networks, and the loss function type is connectionist temporal classification), classical models (such as Transformer-based models); there are also ways to implement this step by using character segmentation, feature extraction, and classifiers based on feature engineering.

[0075] In the following embodiments of this application, the ORC technology involving convolutional neural network (CNN), attention mechanism, and character segmentation algorithm will be introduced as an example.

[0076] In addition, in the ORC technology, after performing the above two major steps, according to actual applications, the post-processing optimization step can also be executed. Post-processing optimization generally includes: semantic correction and structured output. Semantic correction means using natural language processing (NLP) technology to correct the wrong words obtained in Step 2. For example, in the field of medical blood glucose data collection, correcting "blood sugar" to "blood glucose". Structured output means reorganizing the recognition results by paragraphs and tables, which is generally achieved by using a rule engine or a graph neural network.

[0077] 2. Hospital Information System

[0078] The Hospital Information System (HIS) is the core platform of medical informatization, integrating functions such as clinical, management, and operation data, and providing strong assistance for the realization of the full-process digital management of the hospital.

[0079] In the process of treating patients in the hospital, entering the collected patient data into HIS can significantly improve the medical quality, efficiency, and management level. For example, at the clinical diagnosis and treatment level, doctors can improve the accuracy of diagnosis and treatment by referring to the complete data view generated by HIS based on the entered data; the paperless operation and cross-departmental collaboration of HIS can achieve the effect of optimizing the diagnosis and treatment process. At the patient service level, when a patient changes the hospital for treatment, the examination results from other hospitals can be directly reused after being shared by HIS, reducing the patient's expenses and improving the patient's medical experience.

[0080] The above introduces the technical terms involved in this application.

[0081] Currently, in the medical field, traditional blood glucose data collection generally includes two methods: manual entry and Bluetooth transmission. Manual entry typically involves using a standard blood glucose meter to measure the patient's blood glucose level. Medical staff then manually record the measured values ​​on paper forms and manually input them into the Hospital Information System (HIS), a rather cumbersome process. While Bluetooth transmission eliminates the need for manual recording of patient blood glucose values, it requires a specialized Bluetooth-enabled blood glucose meter and a customized interface with the hospital information system for convenient data transmission, making it more expensive.

[0082] Currently, due to limitations in medical infrastructure and usage habits, most hospitals still rely on manual data entry when collecting blood glucose data. This method is inefficient and prone to errors due to human error.

[0083] In view of this, this application proposes an automatic data acquisition method and device, which can solve the problems of low efficiency and high error probability when hospitals collect blood glucose data by manual recording.

[0084] The automatic data acquisition method provided in this application will be described in detail below with reference to the accompanying drawings:

[0085] For example, such as Figure 1 As shown, Figure 1 This application provides an architecture diagram of an automatic data acquisition system 100, which includes a data acquisition module 101, an automatic input module 102, an input control module 103, and a script generation and testing module 104.

[0086] The data acquisition module 101 is used to acquire the data to be collected.

[0087] Optionally, the data to be collected is blood glucose data. Correspondingly, the data acquisition module 101 can specifically acquire the data to be collected based on optical character recognition, obtaining the data from the blood glucose record sheet.

[0088] The automated data entry module 102 is used to enter the data to be collected into the Hospital Information System (HIS) according to the automated data entry script. The automated data entry script is a script pre-written and generated by the script generation and testing module 104. This script can automatically open the HIS and the corresponding blood glucose record of the corresponding patient, and simulate mouse and keyboard operations to automatically enter the corresponding blood glucose data.

[0089] The data entry control module 103 is used to record scripts according to the automated task steps during the process of entering the data to be collected into the HIS, thereby realizing the data entry control function. The automated task step recording script is a script pre-written and generated by the script generation and testing module 104. This script can realize the data entry control function, such as viewing the steps that have been executed and the steps that will be executed, and also provides step pause function, step resume function, and sub-functions to start from a specified step.

[0090] The script generation and testing module 104 is used to generate automated data entry scripts and automated task step recording scripts, and to perform pass / fail tests on these scripts. For example, the script generation and testing module 104 can specifically provide functions for designing automated scripts, including designing step names, key step types (such as selecting a patient, opening a blood glucose record, returning to the patient list, etc.), and key step recognition screenshots (the current automated operation uses a computer vision library, such as OpenCV, to recognize specific click locations). When the task is more complex, it can also provide pure code script writing functionality. After the script design is completed, the script generation and testing module 104 can perform pass / fail tests on the generated automated script.

[0091] For example, such as Figure 2 As shown, Figure 2 The automatic data acquisition method provided in this application includes the following steps:

[0092] S201. Obtain the data to be collected.

[0093] Optionally, the data to be collected can be blood glucose data from the medical field.

[0094] In one possible implementation, when the data to be collected is blood glucose data, the automatic data acquisition device may specifically acquire the data by: acquiring a blood glucose record sheet; and acquiring the data to be collected from the blood glucose record sheet through optical character recognition.

[0095] It is understandable that the blood glucose record sheet mentioned above is a blood glucose record sheet formed by medical staff filling in a blank blood glucose record sheet template after measuring the patient's blood glucose value with a blood glucose meter.

[0096] For example, the data categories recorded on a blood glucose record sheet may include: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value. These data categories can be added or removed according to the hospital's needs in actual application. The template paper style for the blood glucose record sheet can also be adjusted according to actual requirements.

[0097] Furthermore, based on the explanation of optical character recognition technology in the preceding technical terminology section, the automatic data acquisition device obtains the data to be collected from the blood glucose record sheet through optical character recognition, which may include: taking a picture of the blood glucose record sheet to obtain image data of the blood glucose record sheet; and determining the data to be collected from the image data of the blood glucose record sheet based on a preset algorithm. The preset algorithm includes at least one of the following: convolutional neural network, attention mechanism, and character segmentation algorithm. The specific process of the automatic data acquisition device obtaining the data to be collected from the blood glucose record sheet through optical character recognition can be found in S301-S303 below, and will not be repeated here.

[0098] For example, blood glucose record sheets can be photographed using electronic devices such as document scanners or high-speed scanners to obtain image data of the blood glucose record sheets containing characters. In practical applications, other electronic devices with photographing or paper scanning functions can also be used to obtain image data of blood glucose record sheets; this application does not impose specific limitations.

[0099] S202. According to the automated data entry script, the data to be collected is entered into HIS.

[0100] The automated data entry script is a pre-generated script that has undergone passability testing.

[0101] Optionally, the automatic data acquisition device enters the data to be collected into the HIS according to the automated entry script. Specifically, this may include: simulating HIS operation instructions according to the automated entry script; and entering the data to be collected into the HIS according to the HIS operation instructions. The process by which the automatic data acquisition device obtains the data to be collected from the blood glucose record sheet through optical character recognition is described in S401-S402 below, and will not be repeated here.

[0102] It should be noted that the automatic data acquisition device can automatically open the HIS and the corresponding blood glucose item sheet for the corresponding patient based on the automatic entry script, and enter the data to be collected obtained in S201 above into the HIS through simulated keyboard and mouse operation.

[0103] For example, the data to be collected in S201 includes the blood glucose value of patient Zhang San, specifically: "Hospitalization No.: ZY20230715001", "Bed No.: 1406A", "Name: Zhang San", "Blood Glucose Type: Fasting Blood Glucose", "Blood Glucose Value: 6.2 mmol / L", and "Detection Time: 2023-07-18 07:30". At this time, the automated script realizes the entry of patient Zhang San's blood glucose data through steps such as starting HIS, locating patient medical records, and blood glucose data entry. For details, please refer to S402 below, which will not be repeated here.

[0104] S203. During the process of inputting the data to be collected into HIS, the script is recorded according to the automated task steps to realize the data input control function.

[0105] Among them, the automated task step recording script is a pre-generated script that has completed passability testing.

[0106] Optionally, the input control function specifically includes the following sub-functions: step recording function, step pause function, step continue function, specified step start function, and log viewing function.

[0107] For example, when automating the entry of blood glucose data, the automated task step recording script records each operation step in real time and maintains a task execution queue, including: completed steps (e.g., logging into HIS, opening patient medical records), currently pending steps (e.g., selecting blood glucose type, entering values), and subsequent steps that have not yet been executed (e.g., saving data, processing the next record). Furthermore, users can view the following through a visual log panel: successfully executed steps (e.g., marked in green), currently in progress (e.g., marked in yellow), and subsequent pending steps (e.g., marked in gray). This enables step recording and log viewing functions.

[0108] Furthermore, the automated task step recording script can provide multiple clickable control buttons on the automated blood glucose data entry interface, corresponding to functions such as "Pause," "Continue," and "Start Specified Step." Medical staff can then manually click the "Pause" button at any time to pause blood glucose data entry, click the "Continue" button to resume the data entry process, and click the "Start Specified Step" button to skip completed steps or start directly from a specific step in the queue. This enables functions for pausing, resuming, and starting a specified step.

[0109] The following example, using blood glucose data from the medical field as an example, illustrates the situation (using HIS based on a Windows operating system as an example):

[0110] At this point, the automatic data acquisition device has executed the following steps: Start the HIS system → Enter username and password → Search for hospital number ZY20230715001 → Open the blood glucose entry interface. The current pause point is: Selecting blood glucose type (fasting blood glucose). The steps to be executed are: Enter the value 6.2 → Select unit mmol / L → Save → Process the next record (Li Si).

[0111] In response to the above situations, users can choose: 1. Continue the data entry process: Complete the remaining steps directly. 2. Start from the "Enter Value" step: Skip the current blood glucose type selection operation. 3. Complete restart: Clear the record and start from step 1 again.

[0112] The above, using specific blood glucose data examples, illustrates the process of recording scripts according to automated task steps to control the data entry process when inputting data into the HIS.

[0113] In one possible implementation, the data entry control function also includes a blood glucose alarm function. This function generates and displays an alarm message when abnormal blood glucose values ​​are found in the data to be collected. The alarm message can be a pop-up warning window in the visual interface or an audio prompt, informing medical staff that the currently entered patient's blood glucose value exceeds the limit, reminding them to promptly review the data and provide appropriate diagnosis and treatment.

[0114] Referring to the example in S202 above, when the automated task step recording script detects that the blood glucose value being entered exceeds the limit (for example, patient Li Si's blood glucose value is 20 mmol / L), it will automatically pop up an alert window and mark the item that exceeds the limit (for example, mark the blood glucose value of patient Li Si with a red border).

[0115] It should be noted that in the above steps, Python+PyAutoGUI can be used to implement functions such as keyboard and mouse control, adding controls, and alarm pop-ups, and OpenCV can be used for interface element recognition. Alternatively, in practical applications, other code can be used to program the corresponding keyboard and mouse control, adding controls, alarm pop-ups, HIS interface element recognition, etc. The embodiments of this application do not impose specific limitations.

[0116] Based on the above technical solution, this application combines optical character recognition technology with automated scripts to automatically identify key data on paper patient data records, avoiding manual entry errors and reducing data deviations. Utilizing automated scripts to simulate keyboard and mouse operations, it can accurately input patient data into the HIS system. Furthermore, because the input time for a single data entry is reduced from minutes to seconds, this application's technical solution is suitable for batch data entry scenarios. Simultaneously, this application also supports functions such as step pause, specified step start, and data anomaly alarms, adapting to the manual review needs in medical scenarios. Therefore, this application's technical solution achieves efficient and accurate input of patient data in the medical field without requiring additional hardware, significantly improving the automation level and reliability of data collection.

[0117] For example, such as Figure 3 As shown, combined with Figure 2 In the automatic data collection method provided in this application, when the type of data to be collected is blood glucose data, the specific steps for obtaining the data to be collected include:

[0118] S301. Obtain the blood glucose record sheet.

[0119] The blood glucose record sheet is formed by medical staff filling in a blank blood glucose record sheet template after measuring the patient's blood glucose value with a blood glucose meter.

[0120] For example, the data categories recorded on a blood glucose record sheet may include: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value. These data categories can be added or removed according to the hospital's needs in actual application. The template paper style for the blood glucose record sheet can also be adjusted according to actual requirements.

[0121] It is understandable that medical staff use templated blood glucose record sheets to record patients' blood glucose data, which helps the data acquisition device to retrieve the data to be collected from the blood glucose record sheet through optical character recognition in subsequent steps.

[0122] S302. Take a picture of the blood glucose record sheet to obtain the image data of the blood glucose record sheet.

[0123] Optionally, the blood glucose record sheet can be photographed using electronic devices such as document scanners or high-speed scanners to obtain image data of the blood glucose record sheet containing characters. In practical applications, other electronic devices with photographing or paper scanning functions can also be used to obtain image data of the blood glucose record sheet; this application does not impose specific limitations.

[0124] S303. Based on a preset algorithm, determine the data to be collected from the image data of the blood glucose record.

[0125] The preset algorithms may include: convolutional neural networks, attention mechanisms, and character segmentation algorithms.

[0126] Optionally, if the preset algorithm includes convolutional neural networks, attention mechanisms, and character segmentation algorithms, this step can be specifically divided into the following three sub-steps:

[0127] (1) Extract features from the image data of the blood glucose record sheet using a convolutional neural network to obtain character feature data.

[0128] Specifically, convolutional neural networks are responsible for image feature extraction, enabling text detection on blood glucose record image data to extract the features of corresponding characters. For example, they can extract multi-scale texture and edge features from image data and locate text regions (such as bounding boxes for fields like hospital number and blood glucose value).

[0129] In one possible implementation, the automatic data acquisition device extracts features from the image data of the blood glucose record using a convolutional neural network to obtain character feature data, which is then executed using the following formulas one and two:

[0130]

[0131] in, This represents the value of the feature map output by the l-th convolutional layer at position (i,j). This represents the weight of the l-th convolutional kernel at position (m, n). b represents the value of the input image or the feature map of the previous layer at (i+m,j+n). (l) k represents the bias term of the l-th layer. h k represents the height of the convolution kernel. w Indicates the width of the convolution kernel;

[0132]

[0133] Where, p h p represents the height of the pooling window. w The value represents the width of the pooling window, and s represents the step size.

[0134] (2) Based on the attention mechanism and character feature data, determine the target character and perform semantic association on the target character;

[0135] Specifically, the attention mechanism is responsible for character recognition and semantic association to improve recognition accuracy in complex scenarios. For example, the attention mechanism dynamically allocates weights, focusing on the context-related regions of the current character (such as ignoring background noise and strengthening the recognition of the "mmol / L" unit symbol); and establishes associations between fields (such as the logical binding between "blood glucose type: fasting blood glucose" and "blood glucose value: 6.2") to avoid mismatches.

[0136] In one possible implementation, the automatic data acquisition device determines the target character based on an attention mechanism and character feature data, and performs semantic association on the target character, which is done through the following formulas three and four:

[0137]

[0138] Where, α i,j Q represents the attention weight for the i-th query corresponding to the j-th key. i K represents the query vector. i This represents the key vector, and score represents the scoring function;

[0139]

[0140] Among them, C i V represents the context vector at position i. j Represents a value vector.

[0141] (3) Based on the character segmentation algorithm, the target characters after semantic association are separated into single characters to obtain the data to be collected.

[0142] Specifically, character segmentation algorithms are used for single-character separation and for handling characters that are connected or have complex layouts. For example, they can segment connected characters (such as the handwritten "12" in cursive) to ensure that each character is input into the recognition model independently; and when the blood glucose value "5.8" is blurred and connected as "5B" due to the photo being blurry, the character segmentation algorithm can separate it into "5", "8" and the decimal point.

[0143] In one possible implementation, the automatic data acquisition device performs single-character separation on the target characters after semantic association according to a character segmentation algorithm to obtain the data to be acquired, which is then executed using the following formulas five and six:

[0144]

[0145] Where I(x,y) represents the pixel value of the binarized image at coordinates (x,y), H(x) represents the vertical projection histogram, V(y) represents the horizontal projection histogram, w represents the width of the image, and h represents the height of the image;

[0146]

[0147] Where L(x,y) represents the connected component label of pixel (x,y), new_label represents the newly discovered connected component number, and existing_label represents the existing label of adjacent pixels.

[0148] It should be noted that Formulas 1 to 6 above are merely possible examples when the preset algorithm includes convolutional neural networks, attention mechanisms, and character segmentation algorithms, and do not mean that the embodiments of this application can only be implemented using the above formulas. In practical applications, other algorithms and formulas can also be used to implement the function of determining the data to be collected from the image data of the blood glucose record in this step, as needed.

[0149] Based on the above technical solution, the automatic data acquisition method provided in this application embodiment can efficiently and completely obtain patient blood glucose data for HIS entry from paper blood glucose record sheets through OCR technology, so as to facilitate the smooth progress of subsequent automatic data entry steps.

[0150] For example, such as Figure 4 As shown, combined with Figure 3 The automatic data collection method provided in this application involves entering the data to be collected into the HIS according to the automated data entry script, specifically including the following steps:

[0151] S401. Simulate HIS operation instructions based on the automated data entry script.

[0152] Optionally, the HIS operation commands can be keyboard and mouse commands simulated by an automated input script, enabling the operation of the HIS. By simulating the aforementioned keyboard and mouse commands through an automated input script, the automatic data acquisition device can control the HIS loaded on the computer.

[0153] For example, in conjunction with the example in S202 above, HIS operation commands can be "double-click the HIS shortcut on the desktop", "start HIS via the program path", "bring up the search box on the HIS homepage", "type the corresponding characters (patient Zhang San's hospital number) in the search box", "click the corresponding button with the mouse", "enter data parameters in the corresponding item column (such as entering patient Zhang San's blood glucose value)", "save the input by simulating the combination key", etc.

[0154] S402. According to the HIS operation instructions, enter the data to be collected into HIS.

[0155] For example, the following example uses blood glucose data from the medical field as the data to be collected, and combines it with the examples in S202 and S401 above (taking HIS based on a Windows operating system as an example) for specific explanation:

[0156] (1) Start HIS

[0157] Understandably, automated scripts can open the HIS homepage by double-clicking the HIS shortcut on the desktop or by launching it through the program path.

[0158] (2) Locate the patient's medical record

[0159] On the HIS homepage, simulate keyboard operation to bring up the search box, and enter the hospital number in the search box to locate patient Zhang San (i.e., enter "Hospital Number: ZY20230715001" in the example above); after that, click the "Laboratory Report" tab of patient Zhang San by simulating mouse click.

[0160] (3) Blood glucose data entry

[0161] Locate and click the "Blood Glucose Entry" button using image recognition;

[0162] By simulating pressing the down arrow key on the keyboard, select the blood glucose type as "fasting blood glucose";

[0163] Enter the value using SendKeys(“6.2”), simulating pressing the Tab key on the keyboard to jump to the unit field, and enter mmol / L;

[0164] Use the date picker to select the detection time as "2023-07-18 07:30";

[0165] Simulate pressing the corresponding keyboard key combination (e.g., Alt+S) to save the input.

[0166] The above describes the process of inputting the data to be collected into the HIS according to the automated data entry script, using specific blood glucose data examples. It should be noted that in the above steps, Python + PyAutoGUI can be used for keyboard and mouse control, combined with OpenCV for interface element recognition. Alternatively, in practical applications, other code can be used to program the corresponding keyboard and mouse control and HIS interface element recognition functions. This application embodiment does not impose specific limitations.

[0167] Based on the above technical solution, the embodiments of this application can accurately enter patient data into the HIS system through the generated automated entry script, shortening the entry time of a single data entry to the second level, without the need for additional hardware expansion, thus achieving efficient and accurate entry of patient data and significantly improving the automation level and reliability of data collection.

[0168] For example, such as Figure 5 As shown, combined with Figure 4 The automatic data collection method provided in this application also includes the following steps:

[0169] S501. Generate automated data entry scripts and automated task step recording scripts.

[0170] Optionally, the generation of automated input scripts can be achieved using various methods such as action recording based on screen recording tools, configuration through a visual process designer, and script generation based on API interfaces. This application embodiment does not impose specific limitations.

[0171] Optionally, the generation of automated task step recording scripts can be achieved using various methods such as automatic conversion of operation logs, process engine configuration, and intelligent recording and playback. This application embodiment does not impose specific limitations.

[0172] In one possible implementation, this step can be executed by the script generation test module described above, so that the automatic data acquisition device generates an automated data entry script and an automated task step recording script.

[0173] S502. Perform pass / fail tests on the automated data entry script and the automated task step recording script.

[0174] Optionally, the automatic data acquisition device can perform passability tests on the automated input script and the automated task step recording script through various methods such as step-by-step debugging and testing, test environment verification, process integrity testing, and control function testing. This application embodiment does not impose specific limitations.

[0175] In one possible implementation, this step can be executed by the script generation test module described above, so that the automatic data acquisition device can perform pass / fail tests on the automated data entry script and the automated task step recording script.

[0176] Based on the above technical solution, the embodiments of this application can generate automated scripts that provide automated data entry and data entry control functions, thereby entering the data to be collected into the HIS, and can also monitor the data entry process, which can meet the needs of manual review in medical scenarios.

[0177] This application embodiment can divide the automatic data acquisition device into functional modules or functional units according to the above method examples. For example, each function can be divided into a separate functional module or functional unit, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or in software functional modules or functional units. The module or unit division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0178] For example, such as Figure 6 The diagram shown is a possible structural schematic of an automatic data acquisition device according to an embodiment of this application. The automatic data acquisition device 600 includes: an acquisition unit 601 and a processing unit 602.

[0179] The acquisition unit 601 is used to acquire the data to be collected.

[0180] The processing unit 602 is used to input the data to be collected into the hospital information system according to the automated data entry script.

[0181] The processing unit 602 is also used to record scripts according to the automated task steps during the process of inputting the data to be collected into the HIS, so as to realize the data input control function.

[0182] Optionally, the processing unit 602 is also used to simulate HIS operation instructions according to the automated input script.

[0183] Optionally, the processing unit 602 is also used to input the data to be collected into the HIS according to the HIS operation instructions.

[0184] Optionally, the acquisition unit 601 is also used to acquire a blood glucose record sheet; wherein the data categories recorded on the blood glucose record sheet include at least one of the following: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value.

[0185] Optionally, the processing unit 602 is also used to acquire data to be collected from the blood glucose record sheet through optical character recognition.

[0186] Optionally, the acquisition unit 601 is also used to capture images of the blood glucose record sheet and acquire image data of the blood glucose record sheet.

[0187] Optionally, the processing unit 602 is further configured to determine the data to be collected from the image data of the blood glucose record based on a preset algorithm; wherein the preset algorithm includes at least one of the following: convolutional neural network (CNN), attention mechanism, and character segmentation algorithm.

[0188] Optionally, the processing unit 602 is also used to extract features from the image data of the blood glucose record based on the convolutional neural network to obtain character feature data.

[0189] Optionally, the processing unit 602 is also configured to determine the target character based on the attention mechanism and character feature data, and to perform semantic association on the target character.

[0190] Optionally, the processing unit 602 is further configured to perform single-character separation on the target characters after semantic association according to the character segmentation algorithm to obtain the data to be collected.

[0191] Optionally, the processing unit 602 is also configured to execute Formula 1 and Formula 2, which include the following:

[0192]

[0193] in, This represents the value of the feature map output by the l-th convolutional layer at position (i,j). This represents the weight of the l-th convolutional kernel at position (m, n). b represents the value of the input image or the feature map of the previous layer at (i+m,j+n). (y) This represents the bias term of the l-th layer, where l represents the height of the convolution kernel, and k represents the bias term of the l-th layer. w Indicates the width of the convolution kernel;

[0194]

[0195] Where, p h p represents the height of the pooling window. w The value represents the width of the pooling window, and s represents the step size.

[0196] Optionally, the processing unit 602 is also configured to execute Formula 3 and Formula 4, which include the following:

[0197]

[0198] Where, α i,j Q represents the attention weight for the i-th query corresponding to the j-th key.i K represents the query vector. i This represents the key vector, and score represents the scoring function;

[0199]

[0200] Among them, C i V represents the context vector at position i. j Represents a value vector;

[0201] Optionally, the processing unit 602 is also configured to execute Formula 5 and Formula 6, which include the following:

[0202]

[0203] Where I(x,y) represents the pixel value of the binarized image at coordinates (x,y), H(x) represents the vertical projection histogram, V(y) represents the horizontal projection histogram, w represents the width of the image, and h represents the height of the image;

[0204]

[0205] Where L(x,y) represents the connected component label of pixel (x,y), new_label represents the newly discovered connected component number, and existing_label represents the existing label of adjacent pixels.

[0206] Optionally, the processing unit 602 is also used to generate automated data entry scripts and automated task step recording scripts.

[0207] Optionally, the processing unit 602 is also used to perform passability tests on the automated data entry script and the automated task step recording script.

[0208] Optionally, the automatic data acquisition device 600 may also include a storage unit ( Figure 6 (shown in dashed box) The storage unit stores a program or instruction. When the processing unit 601 and the acquisition unit 602 execute the program or instruction, the automatic data acquisition device can perform the automatic data acquisition method described in the above method embodiment.

[0209] also, Figure 6 The technical effects of the automatic data acquisition device can be referred to the technical effects of the automatic data acquisition method described in the above embodiments, and will not be repeated here.

[0210] For example, Figure 7 This is another possible structural diagram of the automatic data acquisition method involved in the above embodiments. For example... Figure 7 As shown, the automatic data acquisition device 700 includes: a processor 702.

[0211] The processor 702 is used to control and manage the operation of the automatic data acquisition device 600, for example, to execute the steps performed by the processing unit 601 and the acquisition unit 602 in the automatic data acquisition device 600, and / or to execute other processes of the technical solution described herein.

[0212] The processor 702 described above can implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the contents of this application. The processor can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the contents of this application. The processor can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0213] Optionally, the automatic data acquisition device 700 may further include a communication interface 703, a memory 701, and a bus 704. The communication interface 703 supports communication between the automatic data acquisition device 700 and other network entities. The memory 701 stores the program code and data of the automatic data acquisition device.

[0214] The memory 701 may be a memory in an automatic data acquisition device. The memory may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk or solid-state drive; the memory may also include a combination of the above types of memory.

[0215] The 704 bus can be an Extended Industry Standard Architecture (EISA) bus, etc. The 704 bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0216] Through the above description of the embodiments, those skilled in the art will clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and module described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0217] This application provides a computer program product containing instructions. When the computer program product is run on the electronic device of this application, it causes the computer to execute the automatic data acquisition method described in the above method embodiment.

[0218] This application also provides a computer-readable storage medium storing instructions. When a computer executes these instructions, the electronic device of this application performs each step of the automatic data acquisition device in the method flow shown in the above method embodiment.

[0219] The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), registers, hard disks, optical fibers, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing, or any other form of computer-readable storage medium in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium may also be a component of the processor. The processor and the storage medium may reside in an application-specific integrated circuit (ASIC). In the embodiments of this application, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0220] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for automatic data acquisition, characterized in that, The method includes: Acquire the data to be collected; According to the automated data entry script, the data to be collected is entered into the Hospital Information System (HIS). During the process of inputting the data to be collected into HIS, a script is recorded according to the automated task steps to realize the data input control function.

2. The method according to claim 1, characterized in that, The step of inputting the data to be collected into the HIS according to the automated input script specifically includes: Simulate HIS operation commands based on the automated data entry script; According to the HIS operation instructions, the data to be collected is entered into the HIS.

3. The method according to claim 2, characterized in that, The data entry control function includes at least one of the following sub-functions: step recording function, step pause function, step continue function, specified step start function, log viewing function, and blood glucose alarm function; wherein, the blood glucose alarm function is used to generate and display alarm information when there is an abnormal blood glucose value in the data to be collected.

4. The method according to claim 3, characterized in that, The type of data to be collected is blood glucose data, and the acquisition of the data to be collected specifically includes: Obtain a blood glucose record; wherein the data categories recorded on the blood glucose record include at least one of the following: hospital number, bed number, patient name, blood glucose type, registration time, and blood glucose value; The data to be collected is obtained from the blood glucose record using optical character recognition.

5. The method according to claim 4, characterized in that, The process of obtaining the data to be collected from the blood glucose record based on optical character recognition specifically includes: Take a picture of the blood glucose record sheet to obtain image data of the blood glucose record sheet; Based on a preset algorithm, the data to be collected is determined from the image data of the blood glucose record; wherein the preset algorithm includes at least one of the following: convolutional neural network, attention mechanism and character segmentation algorithm.

6. The method according to claim 5, characterized in that, When the preset algorithm includes convolutional neural networks, attention mechanisms, and character segmentation algorithms, the data to be collected is determined from the image data of the blood glucose record based on the preset algorithm, specifically including: Character feature data is obtained by extracting features from the image data of the blood glucose record using a convolutional neural network. Based on the attention mechanism and the character feature data, the target character is determined, and semantic association is performed on the target character; According to the character segmentation algorithm, the target character after semantic association is separated into individual characters to obtain the data to be collected.

7. The method according to claim 6, characterized in that, The step of extracting features from the image data of the blood glucose record using a convolutional neural network to obtain character feature data is performed using the following formulas one and two: in, This represents the value of the feature map output by the l-th convolutional layer at position (i,j). This represents the weight of the l-th convolutional kernel at position (m, n). b represents the value of the input image or the feature map of the previous layer at (i+m,j+n). (l) k represents the bias term of the l-th layer. h k represents the height of the convolution kernel. w Indicates the width of the convolution kernel; Where, p h p represents the height of the pooling window. w The width of the pooling window is represented by s, and the step size is represented by s. The step of determining the target character based on the attention mechanism and the character feature data, and performing semantic association on the target character, is executed through the following formulas three and four: Where, α i,j Q represents the attention weight for the i-th query corresponding to the j-th key. i K represents the query vector. i This represents the key vector, and score represents the scoring function; Among them, C i V represents the context vector at position i. j Represents a value vector; The step involves performing single-character separation on the target characters after semantic association according to the character segmentation algorithm to obtain the data to be collected, which is executed using the following formulas five and six: Where I(x,y) represents the pixel value of the binarized image at coordinates (x,y), H(x) represents the vertical projection histogram, V(y) represents the horizontal projection histogram, w represents the width of the image, and h represents the height of the image; Where L(x,y) represents the connected component label of pixel (x,y), new_label represents the newly discovered connected component number, and existing_label represents the existing label of adjacent pixels.

8. The method according to claim 7, characterized in that, Before the data to be collected is entered into the HIS according to the automated data entry script, the method further includes: Generate the automated data entry script and the automated task step recording script; The automated data entry script and the automated task step recording script are subjected to passability testing.

9. An automatic data acquisition system, applied to the automatic data acquisition method according to any one of claims 1-8, characterized in that, The automatic data acquisition system includes: a data acquisition module, an automatic input module, an input control module, and a script generation and testing module; The data acquisition module is used to acquire the data to be collected; The automated data entry module is used to enter the data to be collected into the hospital information system (HIS) according to the automated data entry script. The data entry control module is used to record scripts according to automated task steps during the process of entering the data to be collected into the HIS, thereby realizing the data entry control function of the data to be collected. The script generation and testing module is used to generate the automated data entry script and the automated task step recording script, and to perform passability tests on the automated data entry script and the automated task step recording script.

10. An automatic data acquisition device, characterized in that, The automatic data acquisition device includes: an acquisition unit and a processing unit; The acquisition unit is used to acquire the data to be collected; The processing unit is used to input the data to be collected into the hospital information system (HIS) according to the automated input script. The processing unit is also used to record scripts according to the automated task steps during the process of inputting the data to be collected into the HIS, so as to realize the input control function of the data to be collected.

Images