Electronic sphygmomanometer verification identification method and system based on deep learning

Abstract

The invention discloses an electronic sphygmomanometer verification identification method and system based on deep learning, and relates to the technical field of artificial intelligence, and the method comprises the steps: S1, constructing a deep learning identification model by using an electronic sphygmomanometer image data set; s2, acquiring an identification result of the real-time image data of the electronic sphygmomanometer by utilizing the deep learning identification model based on the real-time image data of the electronic sphygmomanometer; and S3, obtaining the verification identification result of the electronic sphygmomanometer according to the identification result of the real-time image data of the electronic sphygmomanometer. According to the electronic sphygmomanometer verification method, the processes of manual reading, counting and original data and detection report editing in the electronic sphygmomanometer verification process are replaced, the problem that the accuracy rate is low when a traditional machine learning method is used for processing different types of electronic sphygmomanometer readings is solved, and the verification efficiency and the recognition accuracy rate are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a deep learning-based electronic sphygmomanometer calibration and identification method and system. Background Art

[0002] Currently, the calibration of traditional electronic blood pressure monitors still relies on manual reading. This is not only time-consuming and labor-intensive, but also prone to recording errors due to operator fatigue or distraction, which in turn affects test efficiency and accuracy. Faced with the growing demand for calibration and the diversification of device types, traditional manual reading methods can no longer meet the requirements for efficient and accurate measurement and testing.

[0003] Currently, there are still the following major difficulties in using machine learning methods to solve the problem of electronic blood pressure monitor readings: (1) Difficulty in identifying digital display units: Electronic blood pressure monitors usually use seven-segment digital tubes to display numbers. How to accurately distinguish the seven digital tubes belonging to the same number and correctly identify the number based on their lighting combination is one of the technical bottlenecks in traditional methods. However, the recognition accuracy of existing solutions is low and it is difficult to meet the actual application needs; (2) The diversity of devices leads to insufficient recognition adaptability: There are many brands and models of electronic blood pressure monitors, and the display screen designs vary greatly. The display screens of some models may contain a large amount of information unrelated to measurement, such as time, patient number, wearing status, etc. These interfering information significantly increases the difficulty of identifying the valid data area, and traditional methods are difficult to adapt to complex and diverse scenarios.

[0004] To address these challenges, artificial intelligence technologies, particularly deep learning and machine vision, are providing a breakthrough for the intelligent development of electronic blood pressure monitor calibration. Combining the successful experience of machine vision algorithms in these areas, their application to data recognition and recording in electronic blood pressure monitor calibration is a growing trend. Furthermore, the generation of original records and calibration certificates for measurement and testing activities is also a growing trend.

[0005] Therefore, there is an urgent need for an electronic blood pressure monitor calibration and identification method and system based on deep learning to address the deficiencies in the existing technology. Summary of the Invention

[0006] The purpose of the present invention is to propose a deep learning-based electronic blood pressure monitor calibration and identification method and system to replace the manual operation process in the electronic blood pressure monitor calibration process and solve the problem of low accuracy of traditional machine learning methods when processing readings of different categories of electronic blood pressure monitors.

[0007] On the one hand, to achieve the above-mentioned purpose, the present invention provides an electronic blood pressure monitor verification and identification method based on deep learning, comprising the following steps: S1. Build a deep learning recognition model using the electronic blood pressure monitor image dataset; S2. Obtaining a recognition result of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; S3. Obtaining an electronic blood pressure monitor verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure monitor.

[0008] Optionally, the deep learning recognition model is implemented based on the YOLOv11 target detection network.

[0009] Optionally, building a deep learning recognition model using an electronic blood pressure monitor image dataset includes: S1-1. Construct an image dataset of an electronic blood pressure monitor; S1-2, using the image dataset of the electronic blood pressure monitor to train and construct a deep learning coarse recognition model; S1-3, processing the image data set of the electronic blood pressure monitor based on the deep learning coarse recognition model to obtain a digital recognition data set; S1-4, training and constructing a deep learning recognition model based on the digital recognition dataset; S1-5. Acquire the deep learning coarse recognition model and the deep learning fine recognition model as a deep learning recognition model; The image data set of the electronic blood pressure monitor includes image data of a region of interest displayed by the electronic blood pressure monitor, and the region of interest includes a systolic pressure region and a diastolic pressure region.

[0010] Optionally, using the image dataset of the electronic blood pressure monitor for training to construct a deep learning coarse recognition model includes: S1-2-1. Perform data enhancement processing on the image data set of the electronic blood pressure monitor to obtain an enhanced data set of the image data set; S1-2-2. Divide the image dataset into an enhanced data set to obtain an enhanced data training set and an enhanced data verification set; S1-2-3. Using the enhanced data training set as input and the image data corresponding to the enhanced data training set as output, an initial deep learning coarse recognition model is trained based on the YOLOv11 target detection network; S1-2-4. Input the initial deep learning coarse recognition model according to the enhanced data verification set to obtain image data corresponding to the enhanced data verification set; S1-2-5. Determine whether the image data corresponding to the enhanced data verification set are all image data of the area of ​​interest. If so, obtain the initial deep learning coarse recognition model as the deep learning coarse recognition model. Otherwise, use the enhanced data verification set to update the enhanced data training set and return to S1-2-3.

[0011] Optionally, obtaining a recognition result of the real-time image data of the electronic blood pressure monitor by using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor includes: Collect real-time image data of electronic blood pressure monitor; Acquire a digital item of a region of interest of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; The recognition result of the real-time image data of the electronic blood pressure monitor is obtained by performing arrangement processing on the digital items of the region of interest of the real-time image data of the electronic blood pressure monitor.

[0012] Optionally, obtaining a digital item of a region of interest of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor includes: Input the deep learning coarse recognition model according to the real-time image data of the electronic blood pressure monitor to obtain a coarse recognition result of the real-time image data of the electronic blood pressure monitor; Using the coarse recognition result of the electronic blood pressure meter real-time image data, intercepting the electronic blood pressure meter real-time image data to obtain a region of interest of the electronic blood pressure meter real-time image data; The deep learning recognition model is input according to the region of interest of the real-time image data of the electronic blood pressure monitor to obtain a digital item of the region of interest of the real-time image data of the electronic blood pressure monitor.

[0013] Optionally, obtaining the electronic blood pressure monitor verification and recognition result according to the recognition result of the real-time image data of the electronic blood pressure monitor includes: Displaying the recognition result of the real-time image data of the electronic blood pressure monitor to obtain the recognition result; The displayed recognition result based on the recognition result is verified using the blood pressure meter verification procedure to obtain the electronic blood pressure meter verification recognition result.

[0014] On the other hand, to achieve the above-mentioned purpose, the present invention provides an electronic sphygmomanometer verification and identification system based on deep learning, comprising: a model building module, a real-time recognition module and a result acquisition module; The model building module is used to build a deep learning recognition model using the electronic blood pressure monitor image dataset; The real-time recognition module is used to obtain the recognition result of the real-time image data of the electronic blood pressure monitor by using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; The result acquisition module is used to obtain the electronic blood pressure monitor verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure monitor.

[0015] Optionally, the system further includes a model optimization module, which is used to optimize the deep learning recognition model using the electronic blood pressure monitor image data set and the electronic blood pressure monitor image data collected in real time.

[0016] Compared with the closest prior art, the present invention has the following beneficial effects: The present invention applies deep learning to solve problems, and can adapt to electronic blood pressure monitors of different brands and models, realize automatic reading, free up human hands, and improve verification efficiency; the present invention adopts a coarse-to-fine recognition process, first identifying the ROI area to be verified, and then identifying the numbers in the area, which can effectively eliminate various interference information in the captured image and improve recognition accuracy; the present invention self-builds an electronic blood pressure monitor image data set and a seven-segment digital display image data set for electronic blood pressure monitor identification, making the model more adaptable to actual scenarios. In summary, the present invention replaces the process of manual reading, counting, editing raw data and verification certificates during the verification process of electronic blood pressure monitors, and solves the problem of low accuracy of traditional machine learning methods when processing readings of electronic blood pressure monitors of different categories, thereby improving verification efficiency and recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a flow chart of a method for verifying and identifying an electronic blood pressure monitor based on deep learning according to an embodiment of the present invention; Figure 2 Flowchart of model training proposed in an embodiment of the present invention; Figure 3 This is a flow chart of the electronic blood pressure monitor verification and identification proposed in an embodiment of the present invention; Figure 4 This is a structural diagram of an electronic blood pressure monitor calibration and identification system based on deep learning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] The terms used in the embodiments of the present invention are only used to explain the specific embodiments of the present invention and are not intended to limit the present invention.

[0021] The present invention aims to utilize deep learning technology to provide a deep learning-based electronic blood pressure monitor calibration and identification method and system, focusing on solving the core problems of digital display unit identification and valid data area extraction, thereby realizing an efficient and accurate automated calibration process.

[0022] like Figure 1 As shown, an embodiment of the present invention provides an electronic blood pressure monitor verification and identification method based on deep learning, comprising: S1. Build a deep learning recognition model using the electronic blood pressure monitor image dataset; S2. Obtaining a recognition result of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; S3. Obtaining an electronic blood pressure monitor verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure monitor.

[0023] Specifically, such as Figure 2 As shown, an image dataset of an electronic blood pressure monitor is constructed; a deep learning coarse recognition model is trained using the image dataset of the electronic blood pressure monitor, and the function of the model is to accurately determine the region of interest for the calibration of the electronic blood pressure monitor in the image; the trained deep learning coarse recognition model is used to batch process the current image dataset of the electronic blood pressure monitor, and images containing seven-segment digital tube numbers are cut out to form a digital recognition dataset; a deep learning fine recognition model is trained using the digital recognition dataset, and the function of the model is to accurately determine the type, category and position information of all numbers appearing in the image; an image of the target electronic blood pressure monitor for calibration is acquired through collection; the acquired image is sequentially handed over to the deep learning coarse recognition model and the deep learning fine recognition model for processing; the recognition system displays the final recognition result on the operation interface for verification; after recording all data in accordance with the requirements of the blood pressure monitor calibration procedure JJG 692-2010, all recorded data can be imported into the original record template of the data through the detection system so as to issue a calibration certificate.

[0024] This embodiment also includes model optimization. During use, each acquired electronic blood pressure monitor image data is recorded and combined with the existing dataset to enrich the dataset. Furthermore, the newly acquired data is more relevant to daily use scenarios, and the model optimized with this new dataset has higher accuracy in these scenarios.

[0025] There are two cases where there are partial errors in the ROI area: one is that the area is incorrect; the other is that the area is correct but incomplete (especially when the first digit is 1, which can easily exclude the 1). Errors in the ROI area are generally the second case; the first case does not occur.

[0026] To address the above issues, the optimizations during training include: (1) improving the quality of the dataset to ensure that the annotations are tightly and completely enclosed by all digits; (2) using data augmentation, adopting translation, scaling, and mosaic data augmentation strategies to simulate the different positions, sizes, and distances of the target area in the image; (3) during the training process, focusing on samples with recognition errors or low confidence, and increasing the frequency of these samples in subsequent training cycles; (4) fine-tuning the loss function: considering the actual situation, the bounding box regression loss in the total loss function needs special attention, so after multiple training attempts, the weight of the bounding box regression loss is set to 8.5.

[0027] To address the issue of incomplete regions, the algorithm has additional optimizations. After the ROI is coarsely identified, the original recognition result is expanded by an additional 10% in both the left and right directions before being cropped. This method can complete potential missing parts without incorporating interference information. There is no vertical expansion because there are no missing areas in that direction. The deep learning coarse recognition model returns the coordinate information of the target box [x1, y1, x2, y2]. The coordinate information of the cropped area can be calculated. x1-[1.1*(x2-x1)] is the horizontal coordinate of the point on the left side of the cropped box, and x2+[1.1*(x2-x1)] is the horizontal coordinate of the point on the right side of the cropped box. The vertical coordinate remains unchanged.

[0028] The core goal of coarse recognition is to ensure that the recognition boxes in the SYS and DIA areas accurately and completely encompass all digits. During the data preparation phase, various electronic blood pressure monitor images are collected (especially those taken in real-world usage scenarios to verify model performance).

[0029] In terms of data augmentation strategy, setting a rotation angle of degrees = 10, a translation amplitude of translate = 0.2, and a shear strength of shear = 10 can help the model adapt to the diversity brought by shooting angles, target positions, and slight deformations. At the same time, the model will benefit from the rich scene combinations brought by Mosaic data augmentation. This enhancement is disabled in the last 800 rounds of training (parameters obtained after multiple training rounds).

[0030] During the model training phase, each round of learning is performed on batches of augmented training data. After forward propagation, the model calculates a loss based on the predicted results and the ground truth annotations. For practical applications, the weight of the bounding box regression loss is increased to 8.5, focusing on the question of whether the recognition box can contain all digits.

[0031] Validation is performed after each training cycle. Focusing on the accuracy of the bounding boxes, visual inspection confirms that all digits within the SYS and DIA regions are fully enclosed. Quantitative metrics such as mAP50, mAP50-95, precision, recall, and the performance of each component loss on the validation set are also recorded. Any incomplete recognitions, positional shifts, or misidentifications are carefully recorded, and analysis is performed to identify specific brands, lighting conditions, or angles under which errors are more likely to occur.

[0032] The last stage is the adjustment and iteration stage. Based on the feedback from the verification stage, the composition of the dataset is adjusted to increase the frequency of occurrence of these samples in subsequent training cycles.

[0033] The present invention provides a deep learning-based electronic sphygmomanometer calibration and identification method and system. A deep learning recognition model is constructed using an electronic sphygmomanometer image dataset. The system first roughly identifies the coordinates of the systolic blood pressure (SYS) and diastolic blood pressure (DIA) regions in real-time images and expands and intercepts the region of interest (ROI). The system then finely identifies the digital categories and coordinates to combine them into numerical values. Finally, the system automatically generates original records and calibration certificates that comply with regulations, achieving full automation from image acquisition to certificate generation. This addresses the low efficiency of traditional manual readings and the insufficient accuracy of traditional algorithms, improves calibration efficiency and accuracy, and promotes intelligent measurement and testing.

[0034] Furthermore, the deep learning recognition model is implemented based on the YOLOv11 target detection network, specifically: Both the deep learning coarse recognition model and the deep learning fine recognition model are implemented based on the YOLOv11 target detection network.

[0035] As a possible implementation, in the above embodiment, step S1 may specifically include the following steps: S1-1. Construct an image dataset of an electronic blood pressure monitor; S1-2, using the image dataset of the electronic blood pressure monitor to train and construct a deep learning coarse recognition model; S1-3, processing the image data set of the electronic blood pressure monitor based on the deep learning coarse recognition model to obtain a digital recognition data set; S1-4, training and constructing a deep learning recognition model based on the digital recognition dataset; S1-5. Obtain the deep learning coarse recognition model and the deep learning fine recognition model as a deep learning recognition model.

[0036] In this embodiment, the image dataset of the electronic blood pressure monitor includes images of the ROI region displayed by the electronic blood pressure monitor, and is primarily constructed by collecting relevant images from the Internet.

[0037] As a possible implementation, in the above embodiment, step S1-2 may specifically include the following steps: S1-2-1. Perform data enhancement processing on the image data set of the electronic blood pressure monitor to obtain an enhanced data set of the image data set; S1-2-2. Divide the image dataset into an enhanced data set to obtain an enhanced data training set and an enhanced data verification set; S1-2-3. Using the enhanced data training set as input and the image data corresponding to the enhanced data training set as output, an initial deep learning coarse recognition model is trained based on the YOLOv11 target detection network; S1-2-4. Input the initial deep learning coarse recognition model according to the enhanced data verification set to obtain image data corresponding to the enhanced data verification set; S1-2-5. Determine whether the image data corresponding to the enhanced data verification set are all image data of the area of ​​interest. If so, obtain the initial deep learning coarse recognition model as the deep learning coarse recognition model. Otherwise, use the enhanced data verification set to update the enhanced data training set and return to S1-2-3.

[0038] As a possible implementation, in the above embodiment, step S2 may specifically include the following steps: S2-1, collecting real-time image data of the electronic blood pressure monitor; S2-2, obtaining a digital item of a region of interest in the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; S2-3. Performing arrangement processing on the digital items of the region of interest of the real-time image data of the electronic blood pressure monitor to obtain a recognition result of the real-time image data of the electronic blood pressure monitor.

[0039] like Figure 3As shown, a computer camera or other image acquisition module is used to obtain the display of the electronic blood pressure monitor and transmit it to the system; a deep learning coarse recognition model is used to recognize the collected image data, that is, according to the verification procedure JJG 692-2010, the position and coordinate information of the ROI area used for verification are identified; the collected image data is processed according to the coordinate information, and each ROI area is cut out from the original image data; a deep learning fine recognition model is used to recognize the category and coordinate information of the digital items in each ROI area; and according to the coordinate information of the recognized digital items, the recognized digital items are restored to specific numbers.

[0040] For electronic blood pressure measurement, there are two test items. The data required for both tests corresponds to the locations of the SYS and DIA displayed on the electronic blood pressure monitor. Therefore, the coarse recognition ROI is the area displayed on the electronic blood pressure monitor. SYS and DIA are the classification outputs of the coarse recognition object detection network.

[0041] The predicted return contains the bounding box coordinates of the target in the form of [x1, y1, x2, y2], where (x1, y1) is the coordinate of the upper left corner of the bounding box, and (x2, y2) is the coordinate of the lower right corner of the bounding box. Based on this coordinate information, the coarsely identified ROI area is cut out from the original image and used as input for the deep learning fine recognition model. Similarly, in the deep learning fine recognition model, the center point coordinates of each recognized number are calculated based on [x1, y1, x2, y2] and used for number combination. For example, if the deep learning fine recognition model recognizes a 1 and a 2, the result needs to be synthesized based on the center coordinate information to be 21 instead of 12.

[0042] As a possible implementation, in the above embodiment, step S2-2 may specifically include the following steps: S2-2-1. Input the deep learning coarse recognition model according to the real-time image data of the electronic blood pressure monitor to obtain a coarse recognition result of the real-time image data of the electronic blood pressure monitor; S2-2-2. Using the coarse recognition result of the electronic blood pressure monitor real-time image data, intercept the electronic blood pressure monitor real-time image data to obtain a region of interest of the electronic blood pressure monitor real-time image data; The intercepted content will produce different results in different scenarios. There are two test items for the inspection of electronic blood pressure monitors, "static pressure indication error" and "blood pressure indication repeatability verification". The display areas that need to record numbers for the two different items are the areas that display SYS and DIA in the electronic blood pressure monitor.

[0043] For the "blood pressure indication repeatability test," the result to be identified is the number within the SYS and DIA areas. The captured content is the ROI area of ​​the measured sample, which may vary depending on the shooting angle, distance, and type of sample being measured.

[0044] However, the "Static Pressure Indication Error" indicator is displayed differently by different blood pressure monitors. There are three options: displaying it in the SYS area, displaying it in the DIA area, and displaying it in both the SYS and DIA areas. For this indicator, the interception criteria are determined based on the confidence level of the deep learning coarse recognition model's results. If multiple targets are identified in a coarse recognition, the area with the highest confidence level is intercepted, and the remaining targets are discarded. In this indicator, only one image is intercepted by coarse recognition.

[0045] S2-2-3. Input the deep learning recognition model according to the region of interest of the real-time image data of the electronic blood pressure monitor to obtain the digital item of the region of interest of the real-time image data of the electronic blood pressure monitor.

[0046] As a possible implementation, in the above embodiment, step S3 may specifically include the following steps: S3-1, displaying the recognition result of the real-time image data of the electronic blood pressure monitor to obtain a display recognition result; The digital recognition results within each ROI area are displayed in the operation interface of the intelligent recognition system for electronic blood pressure monitor calibration based on deep learning.

[0047] S3-2, based on the displayed recognition result, the recognition result is verified using the blood pressure monitor verification procedure to obtain the electronic blood pressure monitor verification recognition result; According to the operational requirements of the Sphygmomanometer Verification Procedure (JJG 692-2010), after completing all verification items, the read data results are uniformly imported into the technical institution's original record template, automatically generating original records. The metrological verification results are evaluated in accordance with the requirements of the Verification Procedure (JJG 692-2010), and a final verification certificate is generated, which can be directly uploaded to the technical institution's electronic certification system. This embodiment, through linkage with the technical institution's electronic certification system, can directly generate original records used to issue verification certificates.

[0048] Considering the characteristics of measurement work and actual usage scenarios, it is impossible to realize the result recognition of electronic blood pressure monitors through traditional computer vision algorithms. The essence of this embodiment from coarse to fine is to complete the result recognition through a two-stage approach.

[0049] Electronic blood pressure monitor results for measurement purposes rely solely on the numbers displayed in the SYS and DIA areas (heart rate, time, and other information are completely useless and represent interference). Even if traditional algorithms (including deep learning models with digital recognition capabilities) can accurately identify each digit, there are still no efficient and accurate algorithms for subsequent digital filtering. This is primarily due to the fact that the relative positions of non-ROIs and ROIs are not fixed, and there is no absolute logic for distinguishing them.

[0050] Therefore, a two-step approach circumvents the aforementioned algorithmic challenges, employing two distinct deep learning models to gradually complete the metrology result recognition task. First, the SYS and DIA regions are located. The coordinate information contained in the recognition response is used to calculate the position of the entire recognition frame and extract the recognition result from the original image. The extracted result is then used in a deep learning fine recognition model to identify the type of digits. Finally, based on the coordinate information of the digits in the fine recognition, the center coordinates of each digit are calculated, and the result is then determined based on the center coordinates. Furthermore, coarse recognition can help eliminate easily misleading icons on the display screen.

[0051] like Figure 4 As shown, an embodiment of the present invention provides an electronic blood pressure monitor verification and identification system based on deep learning, comprising: a model building module, a real-time identification module and a result acquisition module; The model building module is used to build a deep learning recognition model using the electronic blood pressure monitor image dataset; The model building module collects electronic blood pressure monitor image datasets, combines data augmentation strategies, trains a deep learning recognition model based on the YOLOv11 target detection network, and sets the bounding box regression loss weight to 8.5 to build a model that can accurately locate the SYS and DIA areas.

[0052] The real-time recognition module is used to obtain the recognition result of the real-time image data of the electronic blood pressure monitor by using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; The real-time recognition module first obtains the real-time image data of the electronic blood pressure monitor through the image acquisition unit. The model then identifies the bounding box coordinates [x1, y1, x2, y2] of the SYS and DIA regions and expands the ROI area by 10% pixels in the left and right directions. The category and coordinates of the numbers in the ROI image are then identified, and the coordinates of the center points of the numbers are calculated to form specific values, thereby obtaining the recognition results of the real-time image data.

[0053] The result acquisition module is used to obtain the electronic blood pressure meter verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure meter; The result acquisition module displays the identified blood pressure value on the operation interface. In accordance with the requirements of the JJG 692-2010 verification procedure, it automatically imports the data into the original record template to generate the original record. After completing the metrological verification result determination, a verification certificate is generated, and the certificate can be directly uploaded to the technical organization's electronic certificate system.

[0054] Furthermore, the system also includes a model optimization module, which is used to optimize the deep learning recognition model using the electronic blood pressure monitor image data set and the electronic blood pressure monitor image data collected in real time.

[0055] In addition, the system also has a model optimization module, which records the electronic blood pressure monitor image data collected by the real-time recognition module during use, combines it with the original data set to form a new data set, and optimizes the training of the deep learning recognition model, so that the model's recognition accuracy in daily scenarios continues to improve.

[0056] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0057] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0058] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A deep learning-based electronic blood pressure monitor verification and identification method, characterized in that: include: S1. Build a deep learning recognition model using the electronic blood pressure monitor image dataset; S2. Obtaining a recognition result of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; S3. Obtaining an electronic blood pressure monitor verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure monitor.

2. The electronic sphygmomanometer verification and identification method based on deep learning according to claim 1, characterized in that: The deep learning recognition model is implemented based on the YOLOv11 target detection network.

3. The electronic sphygmomanometer verification and identification method based on deep learning according to claim 2, characterized in that: Building a deep learning recognition model using the electronic blood pressure monitor image dataset includes: S1-1. Construct an image dataset of an electronic blood pressure monitor; S1-2, using the image dataset of the electronic blood pressure monitor to train and construct a deep learning coarse recognition model; S1-3, processing the image data set of the electronic blood pressure monitor based on the deep learning coarse recognition model to obtain a digital recognition data set; S1-4, training and constructing a deep learning recognition model based on the digital recognition dataset; S1-5. Acquire the deep learning coarse recognition model and the deep learning fine recognition model as a deep learning recognition model; The image data set of the electronic blood pressure monitor includes image data of a region of interest displayed by the electronic blood pressure monitor, and the region of interest includes a systolic pressure region and a diastolic pressure region.

4. The electronic sphygmomanometer verification and identification method based on deep learning according to claim 3, characterized in that: Using the image dataset of the electronic blood pressure monitor to train and construct a deep learning coarse recognition model includes: S1-2-1. Perform data enhancement processing on the image data set of the electronic blood pressure monitor to obtain an enhanced data set of the image data set; S1-2-2. Divide the image dataset into an enhanced data set to obtain an enhanced data training set and an enhanced data verification set; S1-2-3. Using the enhanced data training set as input and the image data corresponding to the enhanced data training set as output, an initial deep learning coarse recognition model is trained based on the YOLOv11 target detection network; S1-2-4. Input the initial deep learning coarse recognition model according to the enhanced data verification set to obtain image data corresponding to the enhanced data verification set; S1-2-5. Determine whether the image data corresponding to the enhanced data verification set are all image data of the area of ​​interest. If so, obtain the initial deep learning coarse recognition model as the deep learning coarse recognition model. Otherwise, use the enhanced data verification set to update the enhanced data training set and return to S1-2-3.

5. The electronic sphygmomanometer verification and identification method based on deep learning according to claim 3, characterized in that: The recognition result of the real-time image data of the electronic blood pressure monitor obtained by using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor includes: Collect real-time image data of electronic blood pressure monitor; Acquire a digital item of a region of interest of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; The recognition result of the real-time image data of the electronic blood pressure monitor is obtained by performing arrangement processing on the digital items of the region of interest of the real-time image data of the electronic blood pressure monitor.

6. The electronic sphygmomanometer verification and identification method based on deep learning according to claim 5, characterized in that: Acquiring digital items of the region of interest of the real-time image data of the electronic blood pressure monitor using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor includes: Input the deep learning coarse recognition model according to the real-time image data of the electronic blood pressure monitor to obtain a coarse recognition result of the real-time image data of the electronic blood pressure monitor; Using the coarse recognition result of the electronic blood pressure meter real-time image data, intercepting the electronic blood pressure meter real-time image data to obtain a region of interest of the electronic blood pressure meter real-time image data; The deep learning recognition model is input according to the region of interest of the real-time image data of the electronic blood pressure monitor to obtain a digital item of the region of interest of the real-time image data of the electronic blood pressure monitor.

7. The method for verifying and identifying an electronic sphygmomanometer based on deep learning according to claim 1, characterized in that: Obtaining the electronic blood pressure monitor verification and recognition result according to the recognition result of the real-time image data of the electronic blood pressure monitor includes: Displaying the recognition result of the real-time image data of the electronic blood pressure monitor to obtain the recognition result; The displayed recognition result based on the recognition result is verified using the blood pressure meter verification procedure to obtain the electronic blood pressure meter verification recognition result.

8. A deep learning-based electronic blood pressure monitor verification and identification system, implementing the method according to any one of claims 1 to 7, characterized in that: include: Model building module, real-time recognition module and result acquisition module; The model building module is used to build a deep learning recognition model using the electronic blood pressure monitor image dataset; The real-time recognition module is used to obtain the recognition result of the real-time image data of the electronic blood pressure monitor by using the deep learning recognition model based on the real-time image data of the electronic blood pressure monitor; The result acquisition module is used to obtain the electronic blood pressure monitor verification and recognition result based on the recognition result of the real-time image data of the electronic blood pressure monitor.

9. The electronic sphygmomanometer verification and identification system based on deep learning according to claim 8, characterized in that: The system also includes a model optimization module, which is used to optimize the deep learning recognition model using the electronic blood pressure monitor image data set and the electronic blood pressure monitor image data collected in real time.

Images