Method for identifying and managing user's biometric information based on AI digital human medical guidance system

Abstract

The present invention discloses a method for identifying and managing user biometric information based on an AI digital human medical guidance system, comprising: an information collection device and a server connected to the information collection device by signal, the information collection device comprising a biometric information collection module, an image collection module, a first control module and a first database, the server comprising a second control module and a second database; the above structure and method can improve the accuracy of identification; simplify the initial registration procedure, which not only saves time but also reduces the possibility of manual entry errors; and can limit the doctor's access rights, so that the patient can view the patient's medical history only after verification.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method for identifying and managing user biometric information based on an AI digital human medical guidance system. Background Art

[0002] With the continuous development of science and technology, biometric information recognition technology has been widely used in the fields of security, medical treatment, personal health management, etc. Biometric information recognition is a technology that automatically recognizes and verifies an individual's physiological characteristics (such as fingerprints, irises, facial features, etc.) or behavioral characteristics (such as handwriting, walking posture, etc.). These technologies provide users with a convenient way to authenticate their identities and play an important role in improving security.

[0003] However, traditional biometric identification systems have some limitations. For example, they usually rely on only a single type of biometric information for identity verification, which can lead to a high rate of false positives, especially when the user's physical condition changes, such as illness or injury, which may cause changes in biometric information and affect the accuracy of identification. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a method for identifying and managing user biometric information based on an AI digital human medical guidance system, thereby improving the accuracy of identification.

[0005] The technical solution adopted by the present invention to solve its technical problems is to provide a method for identifying and managing user biometric information based on an AI digital human medical guidance system, including: an information collection device and a server connected to the information collection device by signal, the information collection device including a biometric information collection module, an image collection module, a first control module and a first database, the server including a second control module and a second database;

[0006] The method for identifying and managing user biometric information includes the following steps:

[0007] S1. The patient obtains contact information data by contacting the bio-information collection module;

[0008] S2, the image acquisition module photographs the patient to obtain an image data set, integrates the contact information data and the image data set to form a patient data set and transmits it to the first database;

[0009] S3, the first control module preprocesses the patient data set to obtain a preprocessed data set, the first control module transmits the preprocessed data set to the second database, the second control module calculates the similarity between the preprocessed data set and the patient standard data set preset in the second database, and when the similarity is greater than a first preset threshold, transmits the medical record data set corresponding to the patient standard data set to the first database, and integrates the preprocessed data set into the patient standard data set;

[0010] If the patient is visiting the doctor for the first time, the preprocessed data set is stored in the second database as the patient standard data set;

[0011] S4. When the patient's visit is over, the first control module integrates the patient's medical record data into the medical record data set in the second database, and deletes the medical record data set in the first database.

[0012] As a further improvement of the present invention, the information collection device further includes a medical staff authentication module, and before step S1, further includes step S0: the medical staff is authenticated by the medical staff authentication module and chooses to proceed to step S1 or step S2.

[0013] As a further improvement of the present invention, the specific step of obtaining contact information data by the patient contacting the biometric information collection module in step S1 includes: the patient contacts the biometric information collection module with a finger to obtain contact information data having fingerprint information.

[0014] As a further improvement of the present invention, the specific steps of preprocessing in step S3 include:

[0015] S31, performing brightness correction and color correction on the images in the patient data set;

[0016] S32, performing denoising processing on the images in the patient data set;

[0017] S33, using a posture estimation model to determine key parts of the human body and derive the posture of the human body;

[0018] S34. Extract features of the eyes, nose and mouth of the face.

[0019] As a further improvement of the present invention, the specific steps of performing brightness correction and color correction in step S31 include:

[0020] S311, the first control module detects the histogram of the patient image in the patient data set, and if the histogram is concentrated in low grayscale values, lowers the gamma value;

[0021] If the histogram is concentrated at high grayscale values, increase the gamma value;

[0022] S312. Apply the gray world hypothesis method to adjust the ratios of the red, green, and blue channels of the entire patient image to the same ratio.

[0023] As a further improvement of the present invention, the specific steps of using the posture estimation model to determine the key parts of the human body and derive the posture of the human body in step S33 include:

[0024] S331, substituting the images in the patient data set into the posture estimation model for forward propagation to determine the key parts of the human body, and outputting the coordinates of the key parts;

[0025] S332, drawing key points according to the coordinates of key parts and connecting different key points through connecting lines to obtain a human body posture diagram;

[0026] S333. Substitute the human body posture graph into the deep learning model for posture classification to obtain the human body posture.

[0027] As a further improvement of the present invention, the first control module is preset with a first clarity confidence threshold and a second clarity confidence threshold, each coordinate has confidence level data, and the confidence level data includes a first confidence, a second confidence, and a third confidence;

[0028] The step S331 also includes step S3311: when the clarity of the key part is lower than the first clarity confidence threshold, the coordinates corresponding to the key part are set to the first confidence; when the clarity of the key part is higher than the first clarity confidence threshold and lower than the second clarity confidence threshold, the coordinates corresponding to the key part are set to the second confidence, and the remaining coordinates are set to the third confidence.

[0029] As a further improvement of the present invention, the deep learning model is preset with a plurality of posture labels and a plurality of standard posture data sets corresponding to the posture labels, and the first control module is preset with a posture classification threshold;

[0030] In step S333, the human body posture graph is substituted into the deep learning model for posture classification to obtain the human body posture. The specific steps include: deleting the key points set as the first confidence level, and deleting the connecting lines connected to the key points to obtain the corrected posture graph, and converting the corrected posture graph into the corrected posture graph. Figure 1 1. Perform similarity calculation with the standard posture data set, adjust the similarity calculation ratio of the key point set to the second confidence level and the connecting line connected to the key point to 0.5, and adjust the similarity calculation ratio of the remaining key points and the connecting line connected to the key point to 1;

[0031] When the similarity is greater than the posture classification threshold, the posture label corresponding to the standard posture dataset is selected.

[0032] As a further improvement of the present invention, the key parts include shoulders, elbows, wrists, knees, top of the head and neck.

[0033] As a further improvement of the present invention, the plurality of posture tags are tremor, stiffness, forward leaning, foot damage, hand damage and head and neck damage.

[0034] The beneficial effects of the present invention are at least:

[0035] 1. Multiple authentications of contact biometric information and image data can improve the accuracy of identification.

[0036] 2. For patients who visit the hospital for the first time, their pre-processed data can be automatically stored in the second database as standard data, simplifying the initial registration procedure, which not only saves time but also reduces the possibility of manual entry errors.

[0037] 3. Medical record data is only temporarily transmitted to the first database for consultation assistance when necessary, so that doctors can view medical history. The doctor's access rights can be restricted, and patients can view their medical history only after verification. The patient's medical record data is always kept in the cloud and will not be stored in information collection devices. Even if the information collection device is lost, there is no need to worry about the patient's medical record being exposed. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0039] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0040] Reference Figure 1 , the present invention proposes a method for identifying and managing user biometric information based on an AI digital human medical guidance system, comprising: an information collection device and a server connected to the information collection device by signal, the information collection device comprising a biometric information collection module, an image collection module, a first control module and a first database, the server comprising a second control module and a second database;

[0041] The method for identifying and managing user biometric information includes the following steps:

[0042] S1. The patient obtains contact information data by contacting the bio-information collection module;

[0043] S2, the image acquisition module photographs the patient to obtain an image data set, integrates the contact information data and the image data set to form a patient data set and transmits it to the first database;

[0044] S3, the first control module preprocesses the patient data set to obtain a preprocessed data set, the first control module transmits the preprocessed data set to the second database, the second control module calculates the similarity between the preprocessed data set and the patient standard data set preset in the second database, and when the similarity is greater than a first preset threshold, transmits the medical record data set corresponding to the patient standard data set to the first database, and integrates the preprocessed data set into the patient standard data set;

[0045] If the patient is visiting the doctor for the first time, the preprocessed data set is stored in the second database as the patient standard data set;

[0046] S4. When the patient's visit is over, the first control module integrates the patient's medical record data into the medical record data set in the second database, and deletes the medical record data set in the first database.

[0047] Through multiple authentications of contact biometric information and image data, the accuracy of identification can be improved; for patients visiting the doctor for the first time, their pre-processed data can be automatically stored as standard data in the second database, simplifying the initial registration procedure, which not only saves time but also reduces the possibility of manual entry errors; medical record data is only temporarily transmitted to the first database for consultation assistance when necessary, making it convenient for doctors to review medical history; doctors' access rights can be restricted, and patients can only view their medical history after verification; patients' medical record data is always kept in the cloud and will not be stored in information collection devices. Even if the information collection device is lost, there is no need to worry about the patient's medical record being exposed.

[0048] As a further improvement of the present invention, the information collection device further includes a medical staff authentication module, and before step S1, further includes step S0: the medical staff is authenticated by the medical staff authentication module and chooses to proceed to step S1 or step S2.

[0049] For some patients with arm defects, the doctor can directly skip step S1, and the information collection device will work only after the doctor's certification, which can reduce the working time of the information collection device and increase the service life of the information collection device.

[0050] As a further improvement of the present invention, the specific step of obtaining contact information data by the patient contacting the biometric information collection module in step S1 includes: the patient contacts the biometric information collection module with a finger to obtain contact information data having fingerprint information.

[0051] As a further improvement of the present invention, the specific steps of preprocessing in step S3 include:

[0052] S31, performing brightness correction and color correction on the images in the patient data set;

[0053] S32, performing denoising processing on the images in the patient data set;

[0054] S33, using a posture estimation model to determine key parts of the human body and derive the posture of the human body;

[0055] S34. Extract features of the eyes, nose and mouth of the face.

[0056] Ensure that images maintain consistent visual effects under different lighting conditions and improve the visual quality of images; apply filters to remove noise in images and reduce the impact of irrelevant information on key feature extraction; through posture estimation of continuous frames, the system can capture the changing trends of patient movements and provide more detailed posture analysis; through unified preprocessing operations on all collected image data, the consistency and comparability of the data are ensured to facilitate subsequent data analysis.

[0057] As a further improvement of the present invention, the specific steps of performing brightness correction and color correction in step S31 include:

[0058] S311, the first control module detects the histogram of the patient image in the patient data set, and if the histogram is concentrated in low grayscale values, lowers the gamma value;

[0059] If the histogram is concentrated at high grayscale values, increase the gamma value;

[0060] S312. Apply the gray world hypothesis method to adjust the ratios of the red, green, and blue channels of the entire patient image to the same ratio.

[0061] As a further improvement of the present invention, the specific steps of using the posture estimation model to determine the key parts of the human body and derive the posture of the human body in step S33 include:

[0062] S331, substituting the images in the patient data set into the posture estimation model for forward propagation to determine the key parts of the human body, and outputting the coordinates of the key parts;

[0063] S332, drawing key points according to the coordinates of key parts and connecting different key points through connecting lines to obtain a human body posture diagram;

[0064] S333. Substitute the human body posture graph into the deep learning model for posture classification to obtain the human body posture.

[0065] As a further improvement of the present invention, the first control module is preset with a first clarity confidence threshold and a second clarity confidence threshold, each coordinate has confidence level data, and the confidence level data includes a first confidence, a second confidence, and a third confidence;

[0066] The step S331 also includes step S3311: when the clarity of the key part is lower than the first clarity confidence threshold, the coordinates corresponding to the key part are set to the first confidence; when the clarity of the key part is higher than the first clarity confidence threshold and lower than the second clarity confidence threshold, the coordinates corresponding to the key part are set to the second confidence, and the remaining coordinates are set to the third confidence.

[0067] As a further improvement of the present invention, the deep learning model is preset with a plurality of posture labels and a plurality of standard posture data sets corresponding to the posture labels, and the first control module is preset with a posture classification threshold;

[0068] In step S333, the human body posture graph is substituted into the deep learning model for posture classification to obtain the human body posture. The specific steps include: deleting the key points set as the first confidence level, and deleting the connecting lines connected to the key points to obtain the corrected posture graph, and converting the corrected posture graph into the corrected posture graph. Figure 1 1. Perform similarity calculation with the standard posture data set, adjust the similarity calculation ratio of the key point set to the second confidence level and the connecting line connected to the key point to 0.5, and adjust the similarity calculation ratio of the remaining key points and the connecting line connected to the key point to 1;

[0069] When the similarity is greater than the posture classification threshold, the posture label corresponding to the standard posture dataset is selected.

[0070] For example, if a patient's arm is occluded, the confidence scores of the wrist and elbow key points may be low; in this case, we can choose to delete the key points and connecting lines, thereby inferring the approximate position of the wrist based on the shoulder and other visible key points, which can greatly improve the recognition accuracy.

[0071] As a further improvement of the present invention, the key parts include shoulders, elbows, wrists, knees, top of the head and neck.

[0072] As a further improvement of the present invention, the plurality of posture tags are tremor, stiffness, forward leaning, foot damage, hand damage and head and neck damage.

Claims

1. A method for identifying and managing user biometric information based on an AI digital human medical guidance system, characterized in that: include: An information collection device and a server connected to the information collection device by signal, wherein the information collection device includes a biological information collection module, an image collection module, a first control module and a first database, and the server includes a second control module and a second database; The method for identifying and managing user biometric information includes the following steps: S1. The patient obtains contact information data by contacting the bio-information collection module; S2, the image acquisition module photographs the patient to obtain an image data set, integrates the contact information data and the image data set to form a patient data set and transmits it to the first database; S3, the first control module preprocesses the patient data set to obtain a preprocessed data set, the first control module transmits the preprocessed data set to the second database, the second control module calculates the similarity between the preprocessed data set and the patient standard data set preset in the second database, and when the similarity is greater than a first preset threshold, transmits the medical record data set corresponding to the patient standard data set to the first database, and integrates the preprocessed data set into the patient standard data set; If the patient is visiting the doctor for the first time, the preprocessed data set is stored in the second database as the patient standard data set; S4. When the patient's visit is over, the first control module integrates the patient's medical record data into the medical record data set in the second database, and deletes the medical record data set in the first database.

2. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 1 is characterized in that: The information collection device further includes a medical staff authentication module, and before step S1, further includes step S0: the medical staff is authenticated by the medical staff authentication module and chooses to proceed to step S1 or step S2.

3. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 2 is characterized in that: The specific step of obtaining contact information data by the patient contacting the biometric information collection module in step S1 includes: the patient contacts the biometric information collection module with a finger to obtain contact information data having fingerprint information.

4. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 1 is characterized in that: The specific steps of preprocessing in step S3 include: S31, performing brightness correction and color correction on the images in the patient data set; S32, performing denoising processing on the images in the patient data set; S33, using a posture estimation model to determine key parts of the human body and derive the posture of the human body; S34. Extract features of the eyes, nose and mouth of the face.

5. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 4 is characterized in that: The specific steps of performing brightness correction and color correction in step S31 include: S311, the first control module detects the histogram of the patient image in the patient data set, and if the histogram is concentrated in low grayscale values, lowers the gamma value; If the histogram is concentrated at high grayscale values, increase the gamma value; S312. Apply the gray world hypothesis method to adjust the ratios of the red, green, and blue channels of the entire patient image to the same ratio.

6. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 1 is characterized in that: The specific steps of using the posture estimation model to determine the key parts of the human body and derive the posture of the human body in step S33 include: S331, substituting the images in the patient data set into the posture estimation model for forward propagation to determine the key parts of the human body, and outputting the coordinates of the key parts; S332, drawing key points according to the coordinates of key parts and connecting different key points through connecting lines to obtain a human body posture diagram; S333. Substitute the human body posture graph into the deep learning model for posture classification to obtain the human body posture.

7. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 6 is characterized in that: The first control module is preset with a first clarity confidence threshold and a second clarity confidence threshold, each coordinate has confidence level data, and the confidence level data includes a first confidence, a second confidence, and a third confidence; The step S331 also includes step S3311: when the clarity of the key part is lower than the first clarity confidence threshold, the coordinates corresponding to the key part are set to the first confidence; when the clarity of the key part is higher than the first clarity confidence threshold and lower than the second clarity confidence threshold, the coordinates corresponding to the key part are set to the second confidence, and the remaining coordinates are set to the third confidence.

8. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 7 is characterized in that: The deep learning model is preset with a plurality of posture labels and a plurality of standard posture data sets corresponding to the posture labels, and the first control module is preset with a posture classification threshold; In step S333, the human body posture graph is substituted into the deep learning model for posture classification to obtain the specific steps of the human body posture, including: deleting the key points set to the first confidence level, and deleting the connecting lines connected to the key points to obtain the corrected posture graph, performing similarity calculations on the corrected posture graph and the standard posture data set one by one, adjusting the similarity calculation ratio of the key points set to the second confidence level and the connecting lines connected to the key points to 0.5, and adjusting the similarity calculation ratio of the remaining key points and the connecting lines connected to the key points to 1; When the similarity is greater than the posture classification threshold, the posture label corresponding to the standard posture dataset is selected.

9. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 4 is characterized in that: Key areas include shoulders, elbows, wrists, knees, top of the head, and neck.

10. The method for identifying and managing user biometric information based on the AI ​​digital human medical guidance system according to claim 8, characterized in that: The multiple posture labels are tremor, stiffness, leaning forward, foot damage, hand damage, and head and neck damage.

Images